diff --git a/.github/dependabot.yml b/.github/dependabot.yml
index 3aceaad6..6bace5ad 100644
--- a/.github/dependabot.yml
+++ b/.github/dependabot.yml
@@ -13,6 +13,10 @@ updates:
commit-message:
prefix: "chore(deps)"
groups:
+ study-data-libraries:
+ patterns:
+ - pandas
+ - anndata
# One PR for the lint/format/type toolchain. A ruff bump often reformats many
# files; keeping it grouped and separate from test tooling makes the churn
# reviewable in isolation.
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 57e521ee..3814171f 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -74,6 +74,23 @@ jobs:
ruff check scripts
ruff format --check scripts
+ lineage-install:
+ name: Clean lineage install
+ runs-on: ubuntu-latest
+ steps:
+ - uses: actions/checkout@v4
+ - uses: actions/setup-python@v5
+ with:
+ python-version: "3.12"
+ cache: pip
+ - name: Install the lineage extra in a fresh environment
+ run: |
+ python -m venv /tmp/lineage
+ /tmp/lineage/bin/pip install '.[lineage]' -c envs/lineage-constraints.txt
+ cd /tmp
+ lineage/bin/pip check
+ lineage/bin/python -I -c "import cellquorum, cellrank, palantir, scvelo, pygam"
+
typecheck:
name: Typecheck (engine spine)
runs-on: ubuntu-latest
diff --git a/docker/Dockerfile b/docker/Dockerfile
index 3fb89cfb..57b35fb7 100644
--- a/docker/Dockerfile
+++ b/docker/Dockerfile
@@ -68,11 +68,10 @@ RUN micromamba create -y -n cellquorum-gpu -f /tmp/gpu.conda-lock.yml && microma
# the env yml) because the source is present at /opt/cellquorum by now and the
# yml's relative path would resolve against /tmp inside the image.
RUN micromamba run -n cellquorum-gpu pip install -e .
-# Run the CLI IN the GPU env so runtime capability-detection
-# (compute/router.py: cupy + rapids_singlecell import + CUDA device) can see the
-# GPU. cellquorum-gpu is a superset of cellquorum-core, so every in-process
-# stage still works here. Without this override the GPU image would inherit the
-# cpu ENTRYPOINT and run in cellquorum-core, leaving all GPU acceleration dead.
+COPY --chown=$MAMBA_USER:$MAMBA_USER envs/rapids-worker-requirements.txt /tmp/rapids-worker.txt
+RUN micromamba run -n cellquorum-gpu python -m venv /opt/conda/envs/cellquorum-rapids && \
+ /opt/conda/envs/cellquorum-rapids/bin/pip install -r /tmp/rapids-worker.txt
+ENV CELLQUORUM_RAPIDS_PYTHON=/opt/conda/envs/cellquorum-rapids/bin/python
ENTRYPOINT ["/usr/local/bin/_entrypoint.sh", "micromamba", "run", "-n", "cellquorum-gpu", "cellquorum"]
CMD ["--help"]
diff --git a/docs/benchmarks/pandas-anndata-review/README.md b/docs/benchmarks/pandas-anndata-review/README.md
new file mode 100644
index 00000000..9c21109e
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/README.md
@@ -0,0 +1,167 @@
+# SCC communication ranking across dependency versions
+
+The public paired squamous-cell carcinoma reference contains 18 libraries. Both environments scored the same 54,450 library-specific sender, receiver, ligand and receptor combinations. The median within-library Spearman correlation for magnitude was 0.666. A median of 94 of each library’s top 100 interactions remained in the top 100 (range 87–96). Specificity ranks were more stable, with a median correlation of 0.998 and median top-100 overlap of 100 (range 89–100).
+
+
+
+The left panel includes every matched interaction measurement. A smaller percentile means greater inferred magnitude. Tied scores share their mean percentile. The diagonal marks agreement. On the right, each connected pair represents one library. Sorting tied top-100 scores by sender, receiver, ligand and receptor gives a reproducible cutoff.
+
+## Which dependency changed the rankings?
+
+Each run used the same normalized SCC object, LIANA resource and CellQuorum recipe. Three comparisons independently varied the libraries within the diagnostic environment. The first and last controls preserved every magnitude and specificity score exactly.
+
+| Change | Fixed libraries | Median magnitude ρ | Median shared magnitude top 100 |
+|---|---|---:|---:|
+| AnnData 0.12.10 → 0.13.4 | pandas 2.3.3, LIANA 1.10.0 | 1.000 | 100 |
+| LIANA 1.10.0 → 2.1.1 | pandas 2.3.3, AnnData 0.13.4 | 0.666 | 94 |
+| pandas 2.3.3 → 3.1.1 | AnnData 0.13.4, LIANA 2.1.1 | 1.000 | 100 |
+
+The inferred CD8 T cell HMGB1 to keratinocyte SDC1 interaction entered the top 100 in seven libraries with LIANA 2.1.1, up from four with 1.10.0. Across the same 18 libraries, the three optimally matched Tensor-cell2cell sample factors had Pearson correlations of 0.821, 0.973 and 0.958. Matching maximized total sample-loading correlation across factor permutations.
+
+## What changes in the manuscript figures?
+
+
+
+These are the rendered CellQuorum manuscript panels. Each line connects measurements from the same patient. The original panels scale their vertical axes independently. Lower LIANA ranks indicate stronger inferred interactions. The candidate report used the same full ligand and receptor complexes for the LIANA, NicheNet and MultiNicheNet panels.
+
+In all five patients with both TGFB1–ENG measurements, the tumor had a lower LIANA rank and a higher MultiNicheNet product than the matched normal sample in both environments. The median paired rank difference changed from −0.276 to −0.457. The expression-based MultiNicheNet products and all six NicheNet target weights were identical. Tensor factor 1 declined in six of eight paired patients on baseline and four of eight on the candidate. These factor loadings support different donor-level interpretations. The three regulatory panel sources matched to floating-point precision.
+
+[Plotted sample measurements](scc-panel-measurements.csv) · [Paired differences](scc-publication-donor-differences.csv) · [Seven-panel source comparison](scc-publication-sources.csv)
+
+## Which SCC conclusions survive the new ranking?
+
+The 18 libraries cover ten patients, eight with both tumor and normal tissue. For each sender, receiver, ligand and receptor combination, we compared the within-library magnitude rank percentile in tumor with the percentile in normal tissue from the same patient. Negative differences mean the interaction ranked higher in tumor. A missing score in either tissue excludes that pair. Of 10,469 distinct combinations, 4,031 had a complete pair and 973 had five or more. Among those 973, **801 kept the direction of their median difference and 172 reversed it** (17.7%).
+
+| Finding | LIANA 1.10.0 | LIANA 2.1.1 | Interpretation |
+|---|---:|---:|---|
+| Keratinocyte TGFB1 → fibroblast ENG | Tumor ranked higher in 5/5 | 5/5 | MultiNicheNet expression products also favor tumor in the same five pairs. The raw magnitude-score difference grows, while the median within-library advantage shrinks from 32.9 to 24.6 percentile points. |
+| Matched Tensor factor 1 | Lower in tumor in 6/8 | 4/8 | The majority decline disappears. P2 and P10 cross zero, both with small loadings. |
+| CD8 T-cell HMGB1 → keratinocyte SDC1 | Top 100 in 4 libraries | Top 100 in 7 | P2, P4 and P9 tumor libraries enter the top 100. In both versions only 2/4 complete pairs rank higher in tumor. |
+| Fibroblast COL18A1 → fibroblast ITGB5 | Tumor ranked higher in 2/5 | 4/5 | Median tumor minus normal changes from +24.0 to −47.0 percentile points. |
+| Keratinocyte NAMPT → fibroblast ITGA5–ITGB1 | Tumor ranked higher in 4/6 | 1/6 | Median difference changes from −26.1 to +19.2 percentile points. |
+
+The last two examples have the largest reversals when sorted by the smaller absolute median difference in the two versions. Neither interaction entered any library’s top 100. Across the 973 comparisons, **none met a 5% false-discovery threshold** in either version (two-sided paired sign test, ties excluded, Benjamini–Hochberg correction applied separately per version). These ranks support prioritizing candidate signals for follow-up. They cannot establish ligand binding or a change in signaling activity.
+
+
+
+Each row is one patient. The dashed line marks equal tumor and normal values. Panels A and B use the original quantities plotted in the manuscript panels. Panels C and D use rank percentile points. [Editable donor figure](scc-paired-rank-changes.svg) · [All paired donor differences](scc-paired-rank-donors.csv.gz) · [Interaction summaries and adjusted tests](scc-paired-rank-summary.csv.gz)
+
+## Figure and source data
+
+- [All matched magnitude scores and percentiles (gzip CSV)](scc-magnitude-interactions.csv.gz)
+- [Rank correlations, shared top 100 and maximum score changes for each library and control](scc-library-ranks.csv)
+- [Paired tensor factor correlations](scc-tensor-alignment.csv)
+- [Editable comparison figure](scc-rank-comparison.svg)
+- [SCC reference data and recipe](../reference-studies.md)
+
+## Mouse pancreas lineage after the CytoTRACE 2 repair
+
+On the same 3,696 cells and reference recipe, the repaired wheel completed Palantir,
+stochastic scVelo, CellRank and CytoTRACE 2. The analysis environment used pandas
+3.1.1, AnnData 0.13.4 and NumPy 2.3.5. CytoTRACE 2 used its 19-model ensemble
+with pandas 2.3.3 and NumPy 1.26.4 in a separate CPU worker. Both environments
+passed `pip check`. The CLI took 3 min 34 s and reached 6.48 GiB peak resident
+memory in the Linux/WSL validation run. [Installation instructions](../reference-studies.md#extended-pancreas-lineage-from-a-clean-python-312-install)
+
+| Population | Cells | Median potency |
+|---|---:|---:|
+| Ductal | 916 | 0.300627 |
+| Ngn3 low EP | 262 | 0.196888 |
+| Ngn3 high EP | 642 | 0.083378 |
+| Pre-endocrine | 592 | 0.070243 |
+| Alpha | 481 | 0.081263 |
+| Beta | 591 | 0.048151 |
+| Delta | 70 | 0.077665 |
+| Epsilon | 142 | 0.108630 |
+
+Every potency class agreed with the baseline. The largest absolute change in
+any cell's potency score was 4.19 × 10⁻⁹ or less. All six CellRank fate
+probability columns agreed exactly after aligning cell and fate identifiers.
+The 4,800 fitted values across 24 gene–lineage curves agreed within 10⁻¹⁴.
+[Per-cell scores and classes](pancreas/pancreas-potency-by-cell.csv.gz) ·
+[Population medians](pancreas/pancreas-potency-by-population.csv) ·
+[Numerical and rendered-page comparisons](pancreas/comparison.json)
+
+The two restored publication pages use the same Palantir panel. Their eight
+bar lengths reproduce the median pseudotime from the final analysis object.
+The largest absolute difference in a plotted median was 2.23 × 10⁻¹⁶.
+With Matplotlib 3.11.2 and Pillow 13.0.0, the standalone and assembled pages
+differed from baseline in 1.65% and 3.81% of pixels at 150 dpi. Repeating the
+full run with the baseline renderer (Matplotlib 3.10.9 and Pillow 12.3.0)
+reproduced both pages pixel for pixel and all nine diagnostic PNGs exactly.
+The lineage constraints now pin that renderer.
+[Diagnostic PNG comparisons](pancreas/diagnostic-pngs.csv)
+[Plotted data](pancreas/lineage.palantir.csv) ·
+[Standalone PDF](pancreas/lineage.palantir.pdf) · [Standalone SVG](pancreas/lineage.palantir.svg) ·
+[Assembled PDF](pancreas/lineage.pdf) · [Assembled SVG](pancreas/lineage.svg)
+
+
+
+
+
+The shared lineage embedding selector and figure scales were also exercised on
+the 3,696-cell endocrine pancreas reference. All cell-level potency scores,
+six fate probabilities per cell, and 4,800 fitted expression values across
+24 gene–lineage curves matched the prior candidate. The five trajectory PNGs
+and the two rendered publication PDFs had identical pixels. [Lineage comparison](pancreas/consolidation-check.json)
+
+## GPU workstation validation
+
+A clean install of the Linux GPU lock ran the CellQuorum wheel outside the checkout, with pandas 3.1.1 and AnnData 0.13.4 in the analysis process. The CUDA worker used pandas 3.0.6 with cuDF 26.10.2 and RAPIDS-singlecell 0.18.1. Both environments passed `pip check`. The reference recipe required GPU execution for the routed methods and treated skipped selected analyses as failures.
+
+The input was the public human pancreas reference with 8,391 inDrop cells and 2,961 Smart-seq2 cells, eight published cell types and 1,000 genes. All six selected analysis stages completed and produced the HTML report. The output count layer and cell and gene order matched the input. The graph covered every cell, Leiden produced 17 clusters, and every cell had finite PCA, Harmony, scVI, scANVI and UMAP coordinates.
+
+| Measurement | Earlier reference | New GPU install |
+|---|---:|---:|
+| Cross-study nearest-neighbor cell-type agreement, PCA | 83.95% | 83.95% |
+| Harmony | — | 80.81% |
+| scVI | 81.24% | 81.24% |
+| scANVI | 89.75% | 89.75% |
+
+For each cell we found its closest cell from the other study and compared the published labels. scANVI used those labels during training, so its agreement measures label preservation. The earlier latent-model recipe omitted Harmony. Stage records identify RAPIDS for PCA, Harmony, neighbors/Leiden and UMAP. Lightning recorded CUDA training for scVI, the scANVI pretraining model and scANVI. Input to report took 2 min 8 s on one RTX 4000 Ada. One-second device-wide samples peaked at 42% utilization and 2,619 MiB used, including other processes. An earlier GPU install gave the same four cross-study agreement values and 18 Leiden clusters. Agreement between the two partitions was 0.932 by adjusted Rand index. The UMAP images differed across the two runs. The final install added the missing LIANA orthology dependency, pooch.
+
+All 34 analysis and five RAPIDS package imports passed in their respective interpreters. A second CLI run on the frozen SCC input recomputed LIANA on CPU in this GPU workstation installation. All 54,450 magnitude scores and all 54,450 specificity scores matched the candidate SCC analysis exactly across all 18 libraries. [SCC rerun evidence](gpu/scc-rank-verification.json)
+
+
+
+[Editable figure](gpu/categorical_umap_cell_type.pdf) · [Same coordinates by study](gpu/categorical_umap_study.png) · [Per-cell coordinates, cell type, study and Leiden cluster](gpu/pancreas-umap-source.csv.gz) · [Validated versions, checks and measurements](gpu/validation.json) · [Portable run configuration](gpu/pancreas-gpu.yml)
+
+To reproduce the comparison, follow the GPU installation in `envs/README.md`. Place the [prepared pancreas reference](../reference-studies.md) and the provided YAML in the same directory, then run `cellquorum run -c pancreas-gpu.yml` there. From the repository, run `python scripts/verify_reference_study.py /path/to/runs/pancreas-gpu --data /path/to/input` for the cross-study check.
+
+## CPU workstation validation
+
+A fresh Linux CPU installation ran the published inDrop and Smart-seq2 human
+pancreas counts through preprocessing, feature selection, PCA, Harmony, Leiden,
+the integration benchmark and UMAP. On 4,500 cells sampled from the 11,352-cell
+cohort, Harmony increased iLISI from 0.005 to 0.334. Label agreement (NMI) was
+0.737 before Harmony and 0.738 after it. Graph connectivity was 0.987 in both
+embeddings. With pandas 3, scib-metrics 0.5.9 failed to compute that metric.
+Version 0.5.10 produced all 18 scores. The stage writes them to
+`reports/integration_quality.csv` and links the table from the HTML report.
+The reference check compares the delivered table against the recorded metrics.
+
+A second run stopped after preprocessing and requested a restart at clustering.
+It restored preprocessing and replayed feature selection, PCA and Harmony.
+The counts, normalized expression, coordinates, Leiden assignments and all 18
+scores matched the clean run exactly. All nine PNG panels had identical pixels.
+
+A repeat benchmark on the same final object reused the original measurements.
+Overwriting the score table triggered recomputation and restored the same file hash.
+Permuting the published cell-type labels with seed 65 caused a fresh calculation
+and reduced Harmony NMI to 0.007. In a separate CPU LIANA run on the frozen SCC
+analysis object, all 54,450 rows across 18 libraries matched the candidate
+release in magnitude and specificity rank.
+
+To reproduce the checkpoint comparison, place the prepared human pancreas
+input and the [CPU study configuration](cpu/pancreas-cpu.yml) in the same directory
+and run these commands from that directory.
+
+~~~bash
+cellquorum run -c pancreas-cpu.yml -o runs/clean
+cellquorum run -c pancreas-cpu.yml -o runs/resumed --until-stage preprocessing
+cellquorum run -c pancreas-cpu.yml -o runs/resumed --from-stage clustering
+~~~
+
+The [measurements and file hashes](cpu/validation.json) include pixel comparisons
+for every panel, the maximum difference for each embedding, input and lock hashes,
+the SCC rank comparison, and the benchmark fingerprints before and after shuffling.
diff --git a/docs/benchmarks/pandas-anndata-review/cpu/pancreas-cpu.yml b/docs/benchmarks/pandas-anndata-review/cpu/pancreas-cpu.yml
new file mode 100644
index 00000000..e19a0d99
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/cpu/pancreas-cpu.yml
@@ -0,0 +1,103 @@
+project:
+ name: pancreas_integration
+ organism: human
+paths:
+ output_dir: runs/pancreas-cpu
+input:
+ h5ad: integration.h5ad
+run:
+ profile: publication
+ random_seed: 42
+ require_selected_methods: true
+ continue_on_stage_failure: false
+ checkpoint: true
+ checkpoint_after:
+ - preprocessing
+ resume: true
+stages:
+ ambient_correction: false
+ qc_splice_metrics: false
+ qc_evidence: false
+ preprocessing: true
+ feature_selection: true
+ dimensionality: true
+ integration: true
+ integration_benchmark: true
+ clustering: true
+ annotation: false
+ annotation_diagnostics: false
+ annotation_consensus: false
+ annotation_viz: false
+ reference_mapping: false
+ population_identity: false
+ state_scoring: false
+ state_scoring_viz: false
+ discovery: false
+ subclustering: false
+ query_projection: false
+ qc_finalization: false
+ adjudication: false
+ metacells: false
+ composition: false
+ differential_expression: false
+ differential_state: false
+ coexpression: false
+ grn: false
+ perturbation: false
+ differential_abundance: false
+ enrichment: false
+ enrichment_viz: false
+ de_viz: false
+ ccc_viz: false
+ embeddings: true
+ trajectory: false
+ trajectory_viz: false
+ cell_cell_communication: false
+ network_analysis: false
+ multicellular_programs: false
+ module_remodeling: false
+compute:
+ n_jobs: 2
+ backend: cpu
+ prefer_gpu: false
+ fallback_to_cpu: false
+report:
+ fail_on_report_error: true
+preprocessing:
+ normalization:
+ recipe: cellquorum_log1p_cp10k_v1
+ preserve_counts_layer: counts
+ output_layer: cellquorum_normalized
+feature_selection:
+ counts_layer: counts
+ n_top_genes: 800
+dimensionality:
+ input_layer: cellquorum_normalized
+ n_pcs: 30
+integration:
+ batch_key: study
+ label_key: cell_type
+ n_latent: 15
+ max_epochs: 35
+ random_state: 42
+ methods:
+ - method: harmony
+ output_rep: X_pca_harmony
+integration_benchmark:
+ enabled: true
+ batch_key: study
+ label_key: cell_type
+ mode: full
+ embeddings:
+ - X_pca
+ - X_pca_harmony
+ max_cells: 4500
+clustering:
+ use_rep: X_pca_harmony
+embeddings:
+ embeddings:
+ - umap
+ use_rep: X_pca_harmony
+ color_by:
+ - cell_type
+ - study
diff --git a/docs/benchmarks/pandas-anndata-review/cpu/validation.json b/docs/benchmarks/pandas-anndata-review/cpu/validation.json
new file mode 100644
index 00000000..58307d70
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/cpu/validation.json
@@ -0,0 +1,124 @@
+{
+ "dimensions": [
+ 11352,
+ 1000
+ ],
+ "technologies": {
+ "Pancreas inDrop": 8391,
+ "Pancreas SS2": 2961
+ },
+ "layers_max_difference": {
+ "counts": 0.0,
+ "cellquorum_normalized": 0.0
+ },
+ "coordinates_max_difference": {
+ "X_pca": 0.0,
+ "X_pca_harmony": 0.0,
+ "X_umap": 0.0
+ },
+ "changed_clusters": 0,
+ "metric_count": 18,
+ "integration_metric_max_difference": 0.0,
+ "graph_connectivity": {
+ "X_pca": 0.9869899189581584,
+ "X_pca_harmony": 0.9868237776952048
+ },
+ "biological_checks": {
+ "baseline_iLISI": 0.004922151565551758,
+ "harmony_iLISI": 0.33362698554992676,
+ "baseline_NMI": 0.7368518454720607,
+ "harmony_NMI": 0.7380630887797722
+ },
+ "figures": {
+ "figures/dimensionality_scree.png": {
+ "changed_pixels": 0,
+ "pixels": 516538
+ },
+ "figures/embeddings/categorical_umap_cell_type.png": {
+ "changed_pixels": 0,
+ "pixels": 2697300
+ },
+ "figures/embeddings/categorical_umap_cell_type_paga.png": {
+ "changed_pixels": 0,
+ "pixels": 2697300
+ },
+ "figures/embeddings/categorical_umap_study.png": {
+ "changed_pixels": 0,
+ "pixels": 2652345
+ },
+ "figures/feature_selection_hvg.png": {
+ "changed_pixels": 0,
+ "pixels": 661995
+ },
+ "figures/preprocessing/normalization_depth_correlation_before_after.png": {
+ "changed_pixels": 0,
+ "pixels": 4108955
+ },
+ "figures/preprocessing/normalization_expression_distribution_before_after.png": {
+ "changed_pixels": 0,
+ "pixels": 4165023
+ },
+ "figures/preprocessing/normalization_gene_mean_variance.png": {
+ "changed_pixels": 0,
+ "pixels": 2007792
+ },
+ "figures/preprocessing/normalization_total_counts_before_after.png": {
+ "changed_pixels": 0,
+ "pixels": 4159050
+ }
+ },
+ "scc": {
+ "rows": 54450,
+ "libraries": 18,
+ "max_difference": {
+ "magnitude_rank": 0.0,
+ "specificity_rank": 0.0
+ }
+ },
+ "environment": {
+ "anndata": "0.13.4",
+ "pandas": "3.1.1",
+ "numpy": "2.3.5",
+ "scipy": "1.16.3",
+ "scanpy": "1.12.5",
+ "scib-metrics": "0.5.10",
+ "liana": "2.1.1",
+ "torch": "2.8.0+cpu"
+ },
+ "lock_sha256": "d68cbc3662ce07731417fba8c25278f4518c9d305674bf2817801cbac34c32f9",
+ "input_sha256": {
+ "clean": "8a640f9d4f7c374eb68a3e4115b720f2a404fcfeae8f9e492d64bcd79ddbdcd3",
+ "scc": "0e714ca0a6555c577a14f8992c2cf9dbff852565f2f2211960633233ff8b876c"
+ },
+ "benchmark_cells": 4500,
+ "stage_reuse": {
+ "initial": {
+ "status": "success",
+ "reused": false,
+ "input_fingerprint": "aa9ac748ff01496c1dfee15efc43b5907e40d28eae48a8693f85707f6e45b56b",
+ "harmony_nmi": 0.7380630887797722,
+ "exported_table_sha256": "4d947db03338a973533660ddc61d497e606ad9000a577c33f60d850d2253d5c0"
+ },
+ "repeat": {
+ "status": "skipped",
+ "reused": true,
+ "input_fingerprint": "aa9ac748ff01496c1dfee15efc43b5907e40d28eae48a8693f85707f6e45b56b",
+ "harmony_nmi": 0.7380630887797722,
+ "exported_table_sha256": "4d947db03338a973533660ddc61d497e606ad9000a577c33f60d850d2253d5c0"
+ },
+ "altered": {
+ "status": "success",
+ "reused": false,
+ "input_fingerprint": "aa9ac748ff01496c1dfee15efc43b5907e40d28eae48a8693f85707f6e45b56b",
+ "harmony_nmi": 0.7380630887797722,
+ "exported_table_sha256": "4d947db03338a973533660ddc61d497e606ad9000a577c33f60d850d2253d5c0"
+ },
+ "shuffled": {
+ "status": "success",
+ "reused": false,
+ "input_fingerprint": "2ef2327fccca7c1403f971e0f6c95fef26e92dc5cc2ebf2ffc02c318cf28c957",
+ "harmony_nmi": 0.006993130515903058,
+ "exported_table_sha256": "a6a329cea25cbd007f163cfcf652f4bf1214fb64148d6a5af9fbc43cc2b95c19"
+ }
+ }
+}
diff --git a/docs/benchmarks/pandas-anndata-review/gpu/categorical_umap_cell_type.pdf b/docs/benchmarks/pandas-anndata-review/gpu/categorical_umap_cell_type.pdf
new file mode 100644
index 00000000..8d1b6723
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/gpu/categorical_umap_cell_type.pdf differ
diff --git a/docs/benchmarks/pandas-anndata-review/gpu/categorical_umap_cell_type.png b/docs/benchmarks/pandas-anndata-review/gpu/categorical_umap_cell_type.png
new file mode 100644
index 00000000..38ac167e
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/gpu/categorical_umap_cell_type.png differ
diff --git a/docs/benchmarks/pandas-anndata-review/gpu/categorical_umap_study.png b/docs/benchmarks/pandas-anndata-review/gpu/categorical_umap_study.png
new file mode 100644
index 00000000..434a5aec
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/gpu/categorical_umap_study.png differ
diff --git a/docs/benchmarks/pandas-anndata-review/gpu/pancreas-gpu.yml b/docs/benchmarks/pandas-anndata-review/gpu/pancreas-gpu.yml
new file mode 100644
index 00000000..0eb06ae8
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/gpu/pancreas-gpu.yml
@@ -0,0 +1,104 @@
+project:
+ name: pancreas_latent_sensitivity
+ organism: human
+paths:
+ output_dir: runs/pancreas-gpu
+input:
+ h5ad: integration.h5ad
+run:
+ profile: publication
+ random_seed: 42
+ require_selected_methods: true
+ continue_on_stage_failure: false
+stages:
+ ambient_correction: false
+ qc_splice_metrics: false
+ qc_evidence: false
+ preprocessing: true
+ feature_selection: true
+ dimensionality: true
+ integration: true
+ integration_benchmark: false
+ clustering: true
+ annotation: false
+ annotation_diagnostics: false
+ annotation_consensus: false
+ annotation_viz: false
+ reference_mapping: false
+ population_identity: false
+ state_scoring: false
+ state_scoring_viz: false
+ discovery: false
+ subclustering: false
+ query_projection: false
+ qc_finalization: false
+ adjudication: false
+ metacells: false
+ composition: false
+ differential_expression: false
+ differential_state: false
+ coexpression: false
+ grn: false
+ perturbation: false
+ differential_abundance: false
+ enrichment: false
+ enrichment_viz: false
+ de_viz: false
+ ccc_viz: false
+ embeddings: true
+ trajectory: false
+ trajectory_viz: false
+ cell_cell_communication: false
+ network_analysis: false
+ multicellular_programs: false
+ module_remodeling: false
+compute:
+ n_jobs: 2
+ backend: gpu
+ prefer_gpu: true
+ fallback_to_cpu: false
+report:
+ fail_on_report_error: true
+preprocessing:
+ normalization:
+ recipe: cellquorum_log1p_cp10k_v1
+ preserve_counts_layer: counts
+ output_layer: cellquorum_normalized
+feature_selection:
+ counts_layer: counts
+ n_top_genes: 800
+ method: pearson_residuals
+dimensionality:
+ input_layer: cellquorum_normalized
+ n_pcs: 30
+integration:
+ batch_key: study
+ label_key: cell_type
+ n_latent: 15
+ max_epochs: 35
+ random_state: 42
+ methods:
+ - method: harmony
+ output_rep: X_pca_harmony
+ - method: scvi
+ output_rep: X_scvi
+ - method: scanvi
+ output_rep: X_scanvi
+integration_benchmark:
+ enabled: true
+ batch_key: study
+ label_key: cell_type
+ mode: full
+ embeddings:
+ - X_pca
+ - X_pca_harmony
+ max_cells: 4500
+clustering:
+ use_rep: X_pca_harmony
+embeddings:
+ embeddings:
+ - umap
+ use_rep: X_pca_harmony
+ color_by:
+ - cell_type
+ - study
diff --git a/docs/benchmarks/pandas-anndata-review/gpu/pancreas-umap-source.csv.gz b/docs/benchmarks/pandas-anndata-review/gpu/pancreas-umap-source.csv.gz
new file mode 100644
index 00000000..ad6dd3b9
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/gpu/pancreas-umap-source.csv.gz differ
diff --git a/docs/benchmarks/pandas-anndata-review/gpu/scc-rank-verification.json b/docs/benchmarks/pandas-anndata-review/gpu/scc-rank-verification.json
new file mode 100644
index 00000000..bf3a8103
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/gpu/scc-rank-verification.json
@@ -0,0 +1,21 @@
+{
+ "source": "Frozen normalized SCC reference object",
+ "input_sha256": "0e714ca0a6555c577a14f8992c2cf9dbff852565f2f2211960633233ff8b876c",
+ "pipeline": "installed CellQuorum CLI; LIANA rank_aggregate, CPU",
+ "libraries": 18,
+ "matched_interactions": 54450,
+ "reference_table_sha256": "1dea4350d1d1dc7de3a9e38b6936393db17e1dd7b1708e4d9d0f9be77539af69",
+ "ranking_sha256": "1dea4350d1d1dc7de3a9e38b6936393db17e1dd7b1708e4d9d0f9be77539af69",
+ "magnitude_rank": {
+ "max_abs_diff": 0.0,
+ "exact_matches": 54450,
+ "min_within_library_rank_correlation": 0.9999999999999998
+ },
+ "specificity_rank": {
+ "max_abs_diff": 0.0,
+ "exact_matches": 54450,
+ "min_within_library_rank_correlation": 0.9999999999999998
+ },
+ "cli_complete": "success",
+ "gpu_workstation_lock_sha256": "4014bb5b0d4350560b76ba47cb92e3e198345aee7f274441ac2edbdbb786f9ec"
+}
diff --git a/docs/benchmarks/pandas-anndata-review/gpu/validation.json b/docs/benchmarks/pandas-anndata-review/gpu/validation.json
new file mode 100644
index 00000000..1106a6b0
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/gpu/validation.json
@@ -0,0 +1,79 @@
+{
+ "input_sha256": "8a640f9d4f7c374eb68a3e4115b720f2a404fcfeae8f9e492d64bcd79ddbdcd3",
+ "recipe": "pancreas-gpu.yml",
+ "cells": 11352,
+ "genes": 1000,
+ "studies": {
+ "Pancreas inDrop": 8391,
+ "Pancreas SS2": 2961
+ },
+ "reference_cell_types": 8,
+ "clusters": 17,
+ "input_X_matches_final_counts_and_axis_order": true,
+ "graph_nonempty_for_every_cell": true,
+ "strict_gpu": true,
+ "successful_stages": [
+ "clustering",
+ "dimensionality",
+ "embeddings",
+ "feature_selection",
+ "integration",
+ "preprocessing"
+ ],
+ "routed_compute": {
+ "PCA": "gpu",
+ "Harmony": "gpu",
+ "neighbors_and_Leiden": "gpu",
+ "UMAP": "gpu"
+ },
+ "cuda_training_launches": 3,
+ "controller_packages": {
+ "pandas": "3.1.1",
+ "anndata": "0.13.4",
+ "numpy": "2.3.5",
+ "scipy": "1.16.3",
+ "torch": "2.8.0+cu129",
+ "scvi-tools": "1.5.2",
+ "scanpy": "1.12.5"
+ },
+ "gpu_probe": [
+ "OK: isolated RAPIDS kernel {'cuda_visible': True, 'cuda_device_count': 1, 'device': 'NVIDIA RTX 4000 Ada Generation', 'versions': {'rapids-singlecell-cu12': '0.18.1', 'cudf-cu12': '26.10.2', 'pandas': '3.0.6', 'anndata': '0.13.4'}}",
+ "torch 2.8.0+cu129 cuda_build=12.9 is_built=True is_available=True",
+ "OK: NVIDIA RTX 4000 Ada Generation - CUDA tensor op succeeded.",
+ "OK: scvi-tools importable."
+ ],
+ "pip_check": {
+ "controller": "No broken requirements found.",
+ "worker": "No broken requirements found."
+ },
+ "input_to_report_wall_time": "2:07.74",
+ "max_rss_kib": 2252224,
+ "device_peak_used_mib": 2619,
+ "device_peak_utilization_percent": 42,
+ "device_sampling_interval_seconds": 1,
+ "cross_study_nearest_neighbor_label_agreement": {
+ "X_pca": 0.8394996476391825,
+ "X_scvi": 0.8123678646934461,
+ "X_scanvi": 0.8974630021141649,
+ "X_pca_harmony": 0.8080514446793516
+ },
+ "earlier_reference_label_agreement": {
+ "X_pca": 0.8394996476391825,
+ "X_scvi": 0.8123678646934461,
+ "X_scanvi": 0.8974630021141649
+ },
+ "gpu_lock_sha256": "4014bb5b0d4350560b76ba47cb92e3e198345aee7f274441ac2edbdbb786f9ec",
+ "previous_gpu_run_comparison": {
+ "leiden_adjusted_rand_index": 0.9318026825681062,
+ "clusters_in_runs": [
+ 18,
+ 17
+ ],
+ "umap_procrustes_disparity": 0.0038451434474198244,
+ "cell_type_umap_pixels_identical": false
+ },
+ "import_check": {
+ "controller_passed": 34,
+ "worker_passed": 5
+ }
+}
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/comparison.json b/docs/benchmarks/pandas-anndata-review/pancreas/comparison.json
new file mode 100755
index 00000000..718880e7
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/pancreas/comparison.json
@@ -0,0 +1,16 @@
+{
+ "cells": 3696,
+ "changed_potency_classes": 0,
+ "max_abs_potency_change": 4.180016499599759e-09,
+ "potency_spearman": 0.9999999999999998,
+ "max_abs_pseudotime_change": 1.3322676295501878e-15,
+ "palantir_max_abs_probability_change": 0.0,
+ "cellrank_max_abs_probability_change": 0.0,
+ "fitted_curve_points": 4800,
+ "max_abs_curve_expression_change": 8.43769498715119e-15,
+ "max_abs_panel_median_change": 2.220446049250313e-16,
+ "lineage.palantir_pixel_rms_0_to_255": 0.0,
+ "lineage.palantir_fraction_changed_pixels": 0.0,
+ "lineage_pixel_rms_0_to_255": 0.0,
+ "lineage_fraction_changed_pixels": 0.0
+}
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/consolidation-check.json b/docs/benchmarks/pandas-anndata-review/pancreas/consolidation-check.json
new file mode 100644
index 00000000..b8bc6550
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/pancreas/consolidation-check.json
@@ -0,0 +1,30 @@
+{
+ "cells": 3696,
+ "median_potency": {
+ "Ductal": 0.30062673237806425,
+ "Ngn3 low EP": 0.19688771254527415,
+ "Ngn3 high EP": 0.08337791565194314,
+ "Pre-endocrine": 0.07024311483081029,
+ "Beta": 0.04815135022072501,
+ "Alpha": 0.08126301456645117,
+ "Delta": 0.0776648864533567,
+ "Epsilon": 0.10863032642093637
+ },
+ "fates": 6,
+ "fitted_gene_lineage_curves": 24,
+ "max_potency_difference": 0.0,
+ "max_fate_difference": 0.0,
+ "matched_cell_order": true,
+ "panels_changed_pixels": {
+ "activity_cascade_hallmark.png": 0,
+ "gene_trends.png": 0,
+ "pseudotime_heatmap.png": 0,
+ "velocity_diagnostics.png": 0,
+ "velocity_stream.png": 0,
+ "lineage.pdf": 0,
+ "lineage.palantir.pdf": 0
+ },
+ "trend_rows": 4800,
+ "all_trend_rows_identical": true,
+ "pdf_comparison": "First page rasterized with pdftoppm at a 1600-pixel long edge; both publications are single-page PDFs."
+}
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/diagnostic-pngs.csv b/docs/benchmarks/pandas-anndata-review/pancreas/diagnostic-pngs.csv
new file mode 100755
index 00000000..ea6aa1c2
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/pancreas/diagnostic-pngs.csv
@@ -0,0 +1,10 @@
+figure,same_dimensions,identical_pixels,different_pixel_fraction
+preprocessing/normalization_depth_correlation_before_after.png,True,True,0.0
+preprocessing/normalization_expression_distribution_before_after.png,True,True,0.0
+preprocessing/normalization_gene_mean_variance.png,True,True,0.0
+preprocessing/normalization_total_counts_before_after.png,True,True,0.0
+trajectory/activity_cascade_hallmark.png,True,True,0.0
+trajectory/gene_trends.png,True,True,0.0
+trajectory/pseudotime_heatmap.png,True,True,0.0
+trajectory/velocity_diagnostics.png,True,True,0.0
+trajectory/velocity_stream.png,True,True,0.0
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.csv b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.csv
new file mode 100755
index 00000000..66ef2bfc
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.csv
@@ -0,0 +1,9 @@
+group,value,n_cells
+Ductal,0.04140494689108529,916
+Ngn3 low EP,0.060090180513263494,262
+Ngn3 high EP,0.4055059957991586,642
+Pre-endocrine,0.6514563847037771,592
+Epsilon,0.710440262476072,142
+Beta,0.7906995298835038,591
+Delta,0.7980494100959556,70
+Alpha,0.820060572622252,481
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.pdf b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.pdf
new file mode 100755
index 00000000..369f4ab4
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.pdf differ
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.png b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.png
new file mode 100755
index 00000000..73010d41
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.png differ
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.svg b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.svg
new file mode 100755
index 00000000..3c3d6713
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.palantir.svg
@@ -0,0 +1,311 @@
+
+
+
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/lineage.pdf b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.pdf
new file mode 100755
index 00000000..f752a916
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.pdf differ
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/lineage.png b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.png
new file mode 100755
index 00000000..5fe2203f
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.png differ
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/lineage.svg b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.svg
new file mode 100755
index 00000000..b864669a
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/pancreas/lineage.svg
@@ -0,0 +1,312 @@
+
+
+
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/pancreas-potency-by-cell.csv.gz b/docs/benchmarks/pandas-anndata-review/pancreas/pancreas-potency-by-cell.csv.gz
new file mode 100755
index 00000000..9be3b9a3
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/pancreas/pancreas-potency-by-cell.csv.gz differ
diff --git a/docs/benchmarks/pandas-anndata-review/pancreas/pancreas-potency-by-population.csv b/docs/benchmarks/pandas-anndata-review/pancreas/pancreas-potency-by-population.csv
new file mode 100755
index 00000000..4780954f
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/pancreas/pancreas-potency-by-population.csv
@@ -0,0 +1,9 @@
+population,cells,baseline_median,repaired_median,max_absolute_cell_difference
+Ductal,916,0.30062673286682223,0.30062673237806425,2.5505710876672083e-09
+Ngn3 low EP,262,0.19688771254527626,0.19688771254527415,4.180016499599759e-09
+Epsilon,142,0.1086303264209363,0.10863032642093637,2.991843717214948e-10
+Ngn3 high EP,642,0.08337791565194315,0.08337791565194314,8.956607538967987e-10
+Alpha,481,0.0812630145664514,0.08126301456645117,5.940240799251484e-10
+Delta,70,0.0776648864533569,0.0776648864533567,4.439995315852485e-10
+Pre-endocrine,592,0.0702431148308092,0.07024311483081029,6.327200013922507e-10
+Beta,591,0.0481513502207242,0.04815135022072501,4.824464300279274e-10
diff --git a/docs/benchmarks/pandas-anndata-review/scc-library-ranks.csv b/docs/benchmarks/pandas-anndata-review/scc-library-ranks.csv
new file mode 100644
index 00000000..84546136
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/scc-library-ranks.csv
@@ -0,0 +1,145 @@
+left,right,sample,metric,n_shared,base_only,candidate_only,rho,top100_overlap,max_score_diff
+old,new,P10Normal,magnitude_rank,3266,0,0,0.6699558910365085,94,0.9972451793095692
+old,new,P10Normal,specificity_rank,3266,0,0,0.9956990457834123,100,0.4497824962608077
+old,new,P10Tumor,magnitude_rank,2737,0,0,0.663050669116247,94,0.996712929315046
+old,new,P10Tumor,specificity_rank,2737,0,0,0.9989927871946488,100,0.21152306196674597
+old,new,P1Tumor,magnitude_rank,1811,0,0,0.6681408133025136,91,0.9950331137640445
+old,new,P1Tumor,specificity_rank,1811,0,0,0.9997225803603802,97,0.38161946986647866
+old,new,P2Normal,magnitude_rank,3199,0,0,0.6684950477509963,92,0.997187500113064
+old,new,P2Normal,specificity_rank,3199,0,0,0.998472524861288,100,0.36924525937110897
+old,new,P2Tumor,magnitude_rank,3955,0,0,0.6659185627279876,95,0.9977249747899639
+old,new,P2Tumor,specificity_rank,3955,0,0,0.9985426045722952,100,0.35776009659968167
+old,new,P3Normal,magnitude_rank,1967,0,0,0.6704917310546535,89,0.9954268302428493
+old,new,P3Normal,specificity_rank,1967,0,0,0.999067963092321,97,0.4080542942280694
+old,new,P3Tumor,magnitude_rank,2401,0,0,0.6604679590615412,94,0.996253122806473
+old,new,P3Tumor,specificity_rank,2401,0,0,0.9834484853679981,98,0.7171701039018886
+old,new,P4Tumor,magnitude_rank,4195,0,0,0.665150091134018,96,0.9978551001374436
+old,new,P4Tumor,specificity_rank,4195,0,0,0.9885969091817441,100,0.736099027364489
+old,new,P5Normal,magnitude_rank,1337,0,0,0.6699295921409106,87,0.993273544890253
+old,new,P5Normal,specificity_rank,1337,0,0,0.9963996980073946,89,0.5155490597816939
+old,new,P5Tumor,magnitude_rank,3921,0,0,0.6584496558963988,94,0.9954116754921875
+old,new,P5Tumor,specificity_rank,3921,0,0,0.9950361509517667,100,0.688088328222733
+old,new,P6Normal,magnitude_rank,3363,0,0,0.6673558520403662,94,0.9973246137906566
+old,new,P6Normal,specificity_rank,3363,0,0,0.9988019342605332,100,0.10617924319249372
+old,new,P6Tumor,magnitude_rank,4315,0,0,0.6631970496681772,95,0.9979147359184751
+old,new,P6Tumor,specificity_rank,4315,0,0,0.9968281473276202,100,0.5382171636663265
+old,new,P7Normal,magnitude_rank,3841,0,0,0.670653701211823,93,0.9976574701659472
+old,new,P7Normal,specificity_rank,3841,0,0,0.9967631224878744,99,0.4243271064092853
+old,new,P7Tumor,magnitude_rank,996,0,0,0.6645373116278295,88,0.9909729251556594
+old,new,P7Tumor,specificity_rank,996,0,0,0.9986722449140465,97,0.04877716222023554
+old,new,P8Normal,magnitude_rank,1834,0,0,0.6596779843118681,93,0.9950953690275509
+old,new,P8Normal,specificity_rank,1834,0,0,0.9965586418465807,100,0.4006463227146193
+old,new,P8Tumor,magnitude_rank,3756,0,0,0.6694722165456956,95,0.9976044718084807
+old,new,P8Tumor,specificity_rank,3756,0,0,0.9968728029034284,100,0.6170217367197077
+old,new,P9Normal,magnitude_rank,2518,0,0,0.6653630966465317,90,0.9964271539271267
+old,new,P9Normal,specificity_rank,2518,0,0,0.9986401984603204,97,0.18704612274810017
+old,new,P9Tumor,magnitude_rank,5038,0,0,0.6665285336229274,96,0.9982139313304155
+old,new,P9Tumor,specificity_rank,5038,0,0,0.9988525722220513,100,0.5334783622945896
+old,old-LIANA_pandas2_AnnData013,P10Normal,magnitude_rank,3266,0,0,0.9999999999999999,100,0.0
+old,old-LIANA_pandas2_AnnData013,P10Normal,specificity_rank,3266,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P10Tumor,magnitude_rank,2737,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P10Tumor,specificity_rank,2737,0,0,0.9999999999999999,100,0.0
+old,old-LIANA_pandas2_AnnData013,P1Tumor,magnitude_rank,1811,0,0,0.9999999999999998,100,0.0
+old,old-LIANA_pandas2_AnnData013,P1Tumor,specificity_rank,1811,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P2Normal,magnitude_rank,3199,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P2Normal,specificity_rank,3199,0,0,0.9999999999999999,100,0.0
+old,old-LIANA_pandas2_AnnData013,P2Tumor,magnitude_rank,3955,0,0,0.9999999999999998,100,0.0
+old,old-LIANA_pandas2_AnnData013,P2Tumor,specificity_rank,3955,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P3Normal,magnitude_rank,1967,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P3Normal,specificity_rank,1967,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P3Tumor,magnitude_rank,2401,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P3Tumor,specificity_rank,2401,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P4Tumor,magnitude_rank,4195,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P4Tumor,specificity_rank,4195,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P5Normal,magnitude_rank,1337,0,0,0.9999999999999999,100,0.0
+old,old-LIANA_pandas2_AnnData013,P5Normal,specificity_rank,1337,0,0,0.9999999999999999,100,0.0
+old,old-LIANA_pandas2_AnnData013,P5Tumor,magnitude_rank,3921,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P5Tumor,specificity_rank,3921,0,0,0.9999999999999998,100,0.0
+old,old-LIANA_pandas2_AnnData013,P6Normal,magnitude_rank,3363,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P6Normal,specificity_rank,3363,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P6Tumor,magnitude_rank,4315,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P6Tumor,specificity_rank,4315,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P7Normal,magnitude_rank,3841,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P7Normal,specificity_rank,3841,0,0,0.9999999999999998,100,0.0
+old,old-LIANA_pandas2_AnnData013,P7Tumor,magnitude_rank,996,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P7Tumor,specificity_rank,996,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P8Normal,magnitude_rank,1834,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P8Normal,specificity_rank,1834,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P8Tumor,magnitude_rank,3756,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P8Tumor,specificity_rank,3756,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P9Normal,magnitude_rank,2518,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P9Normal,specificity_rank,2518,0,0,0.9999999999999999,100,0.0
+old,old-LIANA_pandas2_AnnData013,P9Tumor,magnitude_rank,5038,0,0,1.0,100,0.0
+old,old-LIANA_pandas2_AnnData013,P9Tumor,specificity_rank,5038,0,0,1.0,100,0.0
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P10Normal,magnitude_rank,3266,0,0,0.6699558910365085,94,0.9972451793095692
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P10Normal,specificity_rank,3266,0,0,0.9956990457834123,100,0.4497824962608077
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P10Tumor,magnitude_rank,2737,0,0,0.663050669116247,94,0.996712929315046
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P10Tumor,specificity_rank,2737,0,0,0.9989927871946488,100,0.21152306196674597
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P1Tumor,magnitude_rank,1811,0,0,0.6681408133025136,91,0.9950331137640445
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P1Tumor,specificity_rank,1811,0,0,0.9997225803603802,97,0.38161946986647866
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P2Normal,magnitude_rank,3199,0,0,0.6684950477509963,92,0.997187500113064
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P2Normal,specificity_rank,3199,0,0,0.998472524861288,100,0.36924525937110897
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P2Tumor,magnitude_rank,3955,0,0,0.6659185627279876,95,0.9977249747899639
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P2Tumor,specificity_rank,3955,0,0,0.9985426045722952,100,0.35776009659968167
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P3Normal,magnitude_rank,1967,0,0,0.6704917310546535,89,0.9954268302428493
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P3Normal,specificity_rank,1967,0,0,0.999067963092321,97,0.4080542942280694
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P3Tumor,magnitude_rank,2401,0,0,0.6604679590615412,94,0.996253122806473
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P3Tumor,specificity_rank,2401,0,0,0.9834484853679981,98,0.7171701039018886
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P4Tumor,magnitude_rank,4195,0,0,0.665150091134018,96,0.9978551001374436
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P4Tumor,specificity_rank,4195,0,0,0.9885969091817441,100,0.736099027364489
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P5Normal,magnitude_rank,1337,0,0,0.6699295921409106,87,0.993273544890253
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P5Normal,specificity_rank,1337,0,0,0.9963996980073946,89,0.5155490597816939
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P5Tumor,magnitude_rank,3921,0,0,0.6584496558963988,94,0.9954116754921875
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P5Tumor,specificity_rank,3921,0,0,0.9950361509517667,100,0.688088328222733
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P6Normal,magnitude_rank,3363,0,0,0.6673558520403662,94,0.9973246137906566
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P6Normal,specificity_rank,3363,0,0,0.9988019342605332,100,0.10617924319249372
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P6Tumor,magnitude_rank,4315,0,0,0.6631970496681772,95,0.9979147359184751
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P6Tumor,specificity_rank,4315,0,0,0.9968281473276202,100,0.5382171636663265
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P7Normal,magnitude_rank,3841,0,0,0.670653701211823,93,0.9976574701659472
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P7Normal,specificity_rank,3841,0,0,0.9967631224878744,99,0.4243271064092853
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P7Tumor,magnitude_rank,996,0,0,0.6645373116278295,88,0.9909729251556594
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P7Tumor,specificity_rank,996,0,0,0.9986722449140465,97,0.04877716222023554
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P8Normal,magnitude_rank,1834,0,0,0.6596779843118681,93,0.9950953690275509
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P8Normal,specificity_rank,1834,0,0,0.9965586418465807,100,0.4006463227146193
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P8Tumor,magnitude_rank,3756,0,0,0.6694722165456956,95,0.9976044718084807
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P8Tumor,specificity_rank,3756,0,0,0.9968728029034284,100,0.6170217367197077
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P9Normal,magnitude_rank,2518,0,0,0.6653630966465317,90,0.9964271539271267
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P9Normal,specificity_rank,2518,0,0,0.9986401984603204,97,0.18704612274810017
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P9Tumor,magnitude_rank,5038,0,0,0.6665285336229274,96,0.9982139313304155
+old-LIANA_pandas2_AnnData013,new-LIANA_pandas2_AnnData013,P9Tumor,specificity_rank,5038,0,0,0.9988525722220513,100,0.5334783622945896
+new-LIANA_pandas2_AnnData013,new,P10Normal,magnitude_rank,3266,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P10Normal,specificity_rank,3266,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P10Tumor,magnitude_rank,2737,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P10Tumor,specificity_rank,2737,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P1Tumor,magnitude_rank,1811,0,0,0.9999999999999998,100,0.0
+new-LIANA_pandas2_AnnData013,new,P1Tumor,specificity_rank,1811,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P2Normal,magnitude_rank,3199,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P2Normal,specificity_rank,3199,0,0,0.9999999999999999,100,0.0
+new-LIANA_pandas2_AnnData013,new,P2Tumor,magnitude_rank,3955,0,0,0.9999999999999998,100,0.0
+new-LIANA_pandas2_AnnData013,new,P2Tumor,specificity_rank,3955,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P3Normal,magnitude_rank,1967,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P3Normal,specificity_rank,1967,0,0,0.9999999999999998,100,0.0
+new-LIANA_pandas2_AnnData013,new,P3Tumor,magnitude_rank,2401,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P3Tumor,specificity_rank,2401,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P4Tumor,magnitude_rank,4195,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P4Tumor,specificity_rank,4195,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P5Normal,magnitude_rank,1337,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P5Normal,specificity_rank,1337,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P5Tumor,magnitude_rank,3921,0,0,0.9999999999999998,100,0.0
+new-LIANA_pandas2_AnnData013,new,P5Tumor,specificity_rank,3921,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P6Normal,magnitude_rank,3363,0,0,0.9999999999999999,100,0.0
+new-LIANA_pandas2_AnnData013,new,P6Normal,specificity_rank,3363,0,0,0.9999999999999999,100,0.0
+new-LIANA_pandas2_AnnData013,new,P6Tumor,magnitude_rank,4315,0,0,0.9999999999999998,100,0.0
+new-LIANA_pandas2_AnnData013,new,P6Tumor,specificity_rank,4315,0,0,0.9999999999999998,100,0.0
+new-LIANA_pandas2_AnnData013,new,P7Normal,magnitude_rank,3841,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P7Normal,specificity_rank,3841,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P7Tumor,magnitude_rank,996,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P7Tumor,specificity_rank,996,0,0,0.9999999999999998,100,0.0
+new-LIANA_pandas2_AnnData013,new,P8Normal,magnitude_rank,1834,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P8Normal,specificity_rank,1834,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P8Tumor,magnitude_rank,3756,0,0,0.9999999999999998,100,0.0
+new-LIANA_pandas2_AnnData013,new,P8Tumor,specificity_rank,3756,0,0,0.9999999999999998,100,0.0
+new-LIANA_pandas2_AnnData013,new,P9Normal,magnitude_rank,2518,0,0,0.9999999999999999,100,0.0
+new-LIANA_pandas2_AnnData013,new,P9Normal,specificity_rank,2518,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P9Tumor,magnitude_rank,5038,0,0,1.0,100,0.0
+new-LIANA_pandas2_AnnData013,new,P9Tumor,specificity_rank,5038,0,0,0.9999999999999999,100,0.0
diff --git a/docs/benchmarks/pandas-anndata-review/scc-magnitude-interactions.csv.gz b/docs/benchmarks/pandas-anndata-review/scc-magnitude-interactions.csv.gz
new file mode 100644
index 00000000..6f4a8167
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/scc-magnitude-interactions.csv.gz differ
diff --git a/docs/benchmarks/pandas-anndata-review/scc-manuscript-comparison.png b/docs/benchmarks/pandas-anndata-review/scc-manuscript-comparison.png
new file mode 100644
index 00000000..56bcb7ea
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/scc-manuscript-comparison.png differ
diff --git a/docs/benchmarks/pandas-anndata-review/scc-paired-rank-changes.png b/docs/benchmarks/pandas-anndata-review/scc-paired-rank-changes.png
new file mode 100644
index 00000000..841a5a09
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/scc-paired-rank-changes.png differ
diff --git a/docs/benchmarks/pandas-anndata-review/scc-paired-rank-changes.svg b/docs/benchmarks/pandas-anndata-review/scc-paired-rank-changes.svg
new file mode 100644
index 00000000..7722ac93
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/scc-paired-rank-changes.svg
@@ -0,0 +1,900 @@
+
+
+
diff --git a/docs/benchmarks/pandas-anndata-review/scc-paired-rank-donors.csv.gz b/docs/benchmarks/pandas-anndata-review/scc-paired-rank-donors.csv.gz
new file mode 100644
index 00000000..d281f92e
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/scc-paired-rank-donors.csv.gz differ
diff --git a/docs/benchmarks/pandas-anndata-review/scc-paired-rank-summary.csv.gz b/docs/benchmarks/pandas-anndata-review/scc-paired-rank-summary.csv.gz
new file mode 100644
index 00000000..3f815fdd
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/scc-paired-rank-summary.csv.gz differ
diff --git a/docs/benchmarks/pandas-anndata-review/scc-panel-measurements.csv b/docs/benchmarks/pandas-anndata-review/scc-panel-measurements.csv
new file mode 100644
index 00000000..cc412297
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/scc-panel-measurements.csv
@@ -0,0 +1,67 @@
+environment,panel,sample,sender,receiver,ligand,receptor,value,donor,group
+baseline,communication.liana,P10Normal,KC_other,Fibroblast,TGFB1,ENG,1.0,P10,Normal
+baseline,communication.liana,P10Tumor,KC_other,Fibroblast,TGFB1,ENG,0.7239450623379842,P10,Tumor
+baseline,communication.liana,P1Tumor,KC_other,Fibroblast,TGFB1,ENG,0.9489275734049312,P1,Tumor
+baseline,communication.liana,P2Normal,KC_other,Fibroblast,TGFB1,ENG,1.0,P2,Normal
+baseline,communication.liana,P2Tumor,KC_other,Fibroblast,TGFB1,ENG,0.7558908694674163,P2,Tumor
+baseline,communication.liana,P3Tumor,KC_other,Fibroblast,TGFB1,ENG,1.0,P3,Tumor
+baseline,communication.liana,P4Tumor,KC_other,Fibroblast,TGFB1,ENG,0.7060324012035725,P4,Tumor
+baseline,communication.liana,P5Tumor,KC_other,Fibroblast,TGFB1,ENG,0.6633235656148601,P5,Tumor
+baseline,communication.liana,P6Normal,KC_other,Fibroblast,TGFB1,ENG,0.8758693598986214,P6,Normal
+baseline,communication.liana,P6Tumor,KC_other,Fibroblast,TGFB1,ENG,0.482745411389757,P6,Tumor
+baseline,communication.liana,P7Normal,KC_other,Fibroblast,TGFB1,ENG,0.7419590415534121,P7,Normal
+baseline,communication.liana,P8Normal,KC_other,Fibroblast,TGFB1,ENG,1.0,P8,Normal
+baseline,communication.liana,P8Tumor,KC_other,Fibroblast,TGFB1,ENG,0.5425103241702809,P8,Tumor
+baseline,communication.liana,P9Normal,KC_other,Fibroblast,TGFB1,ENG,1.0,P9,Normal
+baseline,communication.liana,P9Tumor,KC_other,Fibroblast,TGFB1,ENG,0.9339662554409132,P9,Tumor
+baseline,communication.tensor,P10Normal,,,,,0.2525872470525308,P10,Normal
+baseline,communication.tensor,P10Tumor,,,,,0.2454410276441849,P10,Tumor
+baseline,communication.tensor,P1Tumor,,,,,0.2413130759909229,P1,Tumor
+baseline,communication.tensor,P2Normal,,,,,0.2108483558079581,P2,Normal
+baseline,communication.tensor,P2Tumor,,,,,0.201705263399212,P2,Tumor
+baseline,communication.tensor,P3Normal,,,,,0.2390301427609089,P3,Normal
+baseline,communication.tensor,P3Tumor,,,,,0.2055838811952062,P3,Tumor
+baseline,communication.tensor,P4Tumor,,,,,0.220874660829927,P4,Tumor
+baseline,communication.tensor,P5Normal,,,,,0.2890546948206635,P5,Normal
+baseline,communication.tensor,P5Tumor,,,,,0.220760233334694,P5,Tumor
+baseline,communication.tensor,P6Normal,,,,,0.2137429157494865,P6,Normal
+baseline,communication.tensor,P6Tumor,,,,,0.2174550602173128,P6,Tumor
+baseline,communication.tensor,P7Normal,,,,,0.2086394668921185,P7,Normal
+baseline,communication.tensor,P7Tumor,,,,,0.2461175786141542,P7,Tumor
+baseline,communication.tensor,P8Normal,,,,,0.3040857683293759,P8,Normal
+baseline,communication.tensor,P8Tumor,,,,,0.2232600678378717,P8,Tumor
+baseline,communication.tensor,P9Normal,,,,,0.2374471034597238,P9,Normal
+baseline,communication.tensor,P9Tumor,,,,,0.237359476114119,P9,Tumor
+candidate,communication.liana,P10Normal,KC_other,Fibroblast,TGFB1,ENG,1.0,P10,Normal
+candidate,communication.liana,P10Tumor,KC_other,Fibroblast,TGFB1,ENG,0.7239451013861387,P10,Tumor
+candidate,communication.liana,P1Tumor,KC_other,Fibroblast,TGFB1,ENG,0.1581157842999113,P1,Tumor
+candidate,communication.liana,P2Normal,KC_other,Fibroblast,TGFB1,ENG,1.0,P2,Normal
+candidate,communication.liana,P2Tumor,KC_other,Fibroblast,TGFB1,ENG,0.7558908588079798,P2,Tumor
+candidate,communication.liana,P3Tumor,KC_other,Fibroblast,TGFB1,ENG,1.0,P3,Tumor
+candidate,communication.liana,P4Tumor,KC_other,Fibroblast,TGFB1,ENG,0.7060324326438197,P4,Tumor
+candidate,communication.liana,P5Tumor,KC_other,Fibroblast,TGFB1,ENG,0.3481181905498559,P5,Tumor
+candidate,communication.liana,P6Normal,KC_other,Fibroblast,TGFB1,ENG,0.8758694198499775,P6,Normal
+candidate,communication.liana,P6Tumor,KC_other,Fibroblast,TGFB1,ENG,0.2850466904568621,P6,Tumor
+candidate,communication.liana,P7Normal,KC_other,Fibroblast,TGFB1,ENG,0.7419590976269921,P7,Normal
+candidate,communication.liana,P8Normal,KC_other,Fibroblast,TGFB1,ENG,1.0,P8,Normal
+candidate,communication.liana,P8Tumor,KC_other,Fibroblast,TGFB1,ENG,0.5425103663716797,P8,Tumor
+candidate,communication.liana,P9Normal,KC_other,Fibroblast,TGFB1,ENG,1.0,P9,Normal
+candidate,communication.liana,P9Tumor,KC_other,Fibroblast,TGFB1,ENG,0.1536625824770459,P9,Tumor
+candidate,communication.tensor,P10Normal,,,,,0.2492320671907798,P10,Normal
+candidate,communication.tensor,P10Tumor,,,,,0.2504150449746599,P10,Tumor
+candidate,communication.tensor,P1Tumor,,,,,0.2460635429111948,P1,Tumor
+candidate,communication.tensor,P2Normal,,,,,0.2122870428543165,P2,Normal
+candidate,communication.tensor,P2Tumor,,,,,0.2150605099192317,P2,Tumor
+candidate,communication.tensor,P3Normal,,,,,0.2227346828009328,P3,Normal
+candidate,communication.tensor,P3Tumor,,,,,0.1901081839786099,P3,Tumor
+candidate,communication.tensor,P4Tumor,,,,,0.2218998044325012,P4,Tumor
+candidate,communication.tensor,P5Normal,,,,,0.2693441809764218,P5,Normal
+candidate,communication.tensor,P5Tumor,,,,,0.2417751033731299,P5,Tumor
+candidate,communication.tensor,P6Normal,,,,,0.2138556637545912,P6,Normal
+candidate,communication.tensor,P6Tumor,,,,,0.2182036100830542,P6,Tumor
+candidate,communication.tensor,P7Normal,,,,,0.2313963117192594,P7,Normal
+candidate,communication.tensor,P7Tumor,,,,,0.2780352867614998,P7,Tumor
+candidate,communication.tensor,P8Normal,,,,,0.2720091924782398,P8,Normal
+candidate,communication.tensor,P8Tumor,,,,,0.2180009153717226,P8,Tumor
+candidate,communication.tensor,P9Normal,,,,,0.2396671861352108,P9,Normal
+candidate,communication.tensor,P9Tumor,,,,,0.2327445745249923,P9,Tumor
diff --git a/docs/benchmarks/pandas-anndata-review/scc-publication-donor-differences.csv b/docs/benchmarks/pandas-anndata-review/scc-publication-donor-differences.csv
new file mode 100644
index 00000000..d80d833c
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/scc-publication-donor-differences.csv
@@ -0,0 +1,35 @@
+donor,baseline_tumor_minus_normal,candidate_tumor_minus_normal,panel
+P10,-0.27605493766201583,-0.27605489861386134,communication.liana
+P2,-0.24410913053258365,-0.2441091411920202,communication.liana
+P6,-0.39312394850886434,-0.5908227293931154,communication.liana
+P8,-0.4574896758297191,-0.4574896336283203,communication.liana
+P9,-0.06603374455908684,-0.8463374175229541,communication.liana
+P10,1.318070982506263,1.318070982506263,communication.mnn
+P2,1.3676857374166191,1.3676857374166191,communication.mnn
+P6,0.062872277669693,0.062872277669693,communication.mnn
+P8,1.312675198190768,1.312675198190768,communication.mnn
+P9,3.1911375480120903,3.1911375480120903,communication.mnn
+P10,-0.007146219408345866,0.0011829777838800737,communication.tensor
+P2,-0.009143092408746112,0.0027734670649152138,communication.tensor
+P3,-0.033446261565702684,-0.03262649882232291,communication.tensor
+P5,-0.06829446148596949,-0.027569077603291914,communication.tensor
+P6,0.003712144467826306,0.004347946328462998,communication.tensor
+P7,0.03747811172203569,0.046638975042240405,communication.tensor
+P8,-0.0808257004915042,-0.054008277106517205,communication.tensor
+P9,-8.762734560480157e-05,-0.006922611610218482,communication.tensor
+P10,-0.013858325974366897,-0.013858325974366897,regulation.scenic
+P2,0.0258452203844695,0.0258452203844695,regulation.scenic
+P3,-0.0198688778040313,-0.0198688778040313,regulation.scenic
+P5,-0.0041458825588518,-0.0041458825588518,regulation.scenic
+P6,0.0009899798722323994,0.0009899798722323994,regulation.scenic
+P8,0.0147785945567515,0.0147785945567515,regulation.scenic
+P9,-0.0517560747936174,-0.0517560747936174,regulation.scenic
+P1,-0.8271685974142287,-0.8271685974142287,regulation.collectri
+P10,0.3444381585279309,0.3444381585279309,regulation.collectri
+P2,-0.5764773537096584,-0.5764773537096584,regulation.collectri
+P3,-0.6262969698389291,-0.6262969698389291,regulation.collectri
+P5,0.745575760570989,0.745575760570989,regulation.collectri
+P6,1.419635678068367,1.419635678068367,regulation.collectri
+P7,0.1302537764246242,0.1302537764246242,regulation.collectri
+P8,-3.5823498301420336,-3.5823498301420336,regulation.collectri
+P9,1.5249170482979877,1.5249170482979877,regulation.collectri
diff --git a/docs/benchmarks/pandas-anndata-review/scc-publication-sources.csv b/docs/benchmarks/pandas-anndata-review/scc-publication-sources.csv
new file mode 100644
index 00000000..53a3587a
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/scc-publication-sources.csv
@@ -0,0 +1,8 @@
+panel,baseline,candidate,shared,max_delta,pearson
+communication.liana,15,15,15,0.7908117891050198,0.48069340640131697
+communication.mnn,10,10,10,0.0,0.9999999999999999
+communication.targets,6,6,6,0.0,1.0
+communication.tensor,18,18,18,0.0320765758511361,0.8213054366355064
+regulation.collectri,19,19,19,8.326672684688674e-17,1.0
+regulation.scenic,15,15,15,0.0,1.0
+regulation.targets,8,8,8,0.0,1.0
diff --git a/docs/benchmarks/pandas-anndata-review/scc-rank-comparison.png b/docs/benchmarks/pandas-anndata-review/scc-rank-comparison.png
new file mode 100644
index 00000000..4017077e
Binary files /dev/null and b/docs/benchmarks/pandas-anndata-review/scc-rank-comparison.png differ
diff --git a/docs/benchmarks/pandas-anndata-review/scc-rank-comparison.svg b/docs/benchmarks/pandas-anndata-review/scc-rank-comparison.svg
new file mode 100644
index 00000000..9ee5348a
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/scc-rank-comparison.svg
@@ -0,0 +1,1100 @@
+
+
+
diff --git a/docs/benchmarks/pandas-anndata-review/scc-tensor-alignment.csv b/docs/benchmarks/pandas-anndata-review/scc-tensor-alignment.csv
new file mode 100644
index 00000000..a273bbeb
--- /dev/null
+++ b/docs/benchmarks/pandas-anndata-review/scc-tensor-alignment.csv
@@ -0,0 +1,4 @@
+baseline,candidate,sample_loading_Pearson_r
+Factor 1,Factor 1,0.8213054366355066
+Factor 2,Factor 2,0.9728306233631888
+Factor 3,Factor 3,0.9583988350409338
diff --git a/docs/benchmarks/reference-studies.md b/docs/benchmarks/reference-studies.md
index 7137c235..fb5d8e16 100644
--- a/docs/benchmarks/reference-studies.md
+++ b/docs/benchmarks/reference-studies.md
@@ -30,7 +30,7 @@ The public-study catalog contains 41 stages and 82 method registrations. Every e
For the two Kang supplements, generate the specimen-labelled input once with `python "$recipes/helpers/prepare_kang_specimens.py" runs/kang_paired_analysis/objects/final_annotated.h5ad kang_sensitivity_input.h5ad`. For SCC, run the paired signaling recipe first. Both network and program recipes use the published four-population slice. After the program run, `python "$recipes/helpers/figure_scc_programs.py" runs/scc_programs` exports the editable four-panel figure, individual source tables and legend in `reports/publication/scc_programs/`.
-For the extended lineage recipe, install the `lineage` extra with the shipped `lineage-constraints.txt` into the current analysis environment and provision the CytoTRACE2 models. The reference used scVelo 0.3.4 with NumPy 2.3.5. Install the standard CellOracle environment including loompy and pyarrow for its separate perturbation run. The SCC program run needs memory for both the Python parent and its R worker. In a 10 GiB container two stability resamples exhausted the cgroup limit. The 14 GiB container completed the five resamples. Strict execution treats a killed resample as an incomplete requested analysis. A whole-VM restart interrupts live computation. The file `logs/stage_events.jsonl` records stage boundaries per attempt. Use a persistent output volume and configured checkpoints for resume. In the endocrine reference, resuming at `trajectory_viz` from the saved trajectory checkpoint reproduced the clean run’s potency scores, pseudotimes, fate probabilities and 4,800 fitted curve coordinates exactly.
+For the extended lineage recipe, follow the [clean lineage and CytoTRACE 2 installation](#extended-pancreas-lineage-from-a-clean-python-312-install) and provision the model ensemble. The reference used scVelo 0.3.4 with NumPy 2.3.5. Install the standard CellOracle environment including loompy and pyarrow for its separate perturbation run. The SCC program run needs memory for both the Python parent and its R worker. In a 10 GiB container two stability resamples exhausted the cgroup limit. The 14 GiB container completed the five resamples. Strict execution treats a killed resample as an incomplete requested analysis. A whole-VM restart interrupts live computation. The file `logs/stage_events.jsonl` records stage boundaries per attempt. Use a persistent output volume and configured checkpoints for resume. In the endocrine reference, resuming at `trajectory_viz` from the saved trajectory checkpoint reproduced the clean run’s potency scores, pseudotimes, fate probabilities and 4,800 fitted curve coordinates exactly.

@@ -58,3 +58,31 @@ An independent wet-lab handoff is outstanding. Give the researcher the [quicksta
4. Retrieve its editable panel, exact source table, legend and resolved settings. Check one plotted value against the table.
A maintainer or automated fixture can verify the files and numbers. Record completion of the human handoff only after an independent researcher performs those tasks.
+
+## Extended pancreas lineage from a clean Python 3.12 install
+
+Run these commands from the source checkout on Linux or WSL. The lineage
+constraints select the tested pandas and AnnData pair, NumPy for stochastic
+velocity, and the upstream PyGPCCA revision compatible with the report renderer.
+Git is needed to install that revision.
+
+```bash
+python3.12 -m venv .venvs/lineage
+.venvs/lineage/bin/python -m pip install '.[lineage]' -c envs/lineage-constraints.txt
+.venvs/lineage/bin/python -m pip check
+
+python3.12 -m venv .venvs/cytotrace2
+.venvs/cytotrace2/bin/python -m pip install -r envs/cytotrace2-requirements.txt
+.venvs/cytotrace2/bin/python -m pip check
+export CELLQUORUM_CYTOTRACE2_PYTHON="$PWD/.venvs/cytotrace2/bin/python"
+export PATH="$PWD/.venvs/lineage/bin:$PATH"
+```
+
+CytoTRACE 2 1.1.0.4 needs NumPy 1 and pandas 2 for the upstream gene mapping.
+CellQuorum passes counts and original barcodes to the worker and records its
+package versions with the per-cell results. On first use the model downloads its
+19 weights. For offline reuse, copy the complete model1.pt through model19.pt
+ensemble from a matching installation into the worker's
+`cytotrace2_py/resources/models` directory. The lineage constraints and worker
+requirements also ship under `share/cellquorum/benchmarks` in the installed
+wheel. Use the same exported interpreter for `cellquorum plan` and `cellquorum run`.
diff --git a/docs/guide/backends.md b/docs/guide/backends.md
index 60e08217..e6f6a207 100644
--- a/docs/guide/backends.md
+++ b/docs/guide/backends.md
@@ -17,7 +17,7 @@ commands, dependency-isolation rationales, and `make lock` usage, see
|---|---|
| `cellquorum-core` | The main runtime (scanpy, anndata ≥ 0.11, scvi-tools, …). The `cellquorum` CLI is invoked from here. |
| `cellquorum-dev` | `cellquorum-core` plus testing/linting/build tooling (pytest, ruff, mypy, docs). |
-| `cellquorum-gpu` | `cellquorum-core` plus CUDA PyTorch, scvi-tools, and the RAPIDS stack for GPU acceleration. |
+| `cellquorum-gpu` | `cellquorum-core` with CUDA PyTorch and scvi-tools. An isolated RAPIDS worker runs with its supported pandas version. Set `CELLQUORUM_RAPIDS_PYTHON` to its Python executable. |
| `cellquorum-r` | The R bridge: Seurat, zellkonverter, rpy2, anndata2ri for R/Bioconductor methods. |
## Isolated backend environments
diff --git a/docs/guide/configuration.md b/docs/guide/configuration.md
index 3bd61d2e..a6277225 100644
--- a/docs/guide/configuration.md
+++ b/docs/guide/configuration.md
@@ -41,6 +41,19 @@ failures separately. In CLI JSON, `disabled_stages` identifies stages switched o
`reused_stages` lists stages whose results came from a validated checkpoint.
The legacy `skipped_stages` field includes all skipped records.
+With `--from-stage`, CellQuorum loads the latest compatible checkpoint before
+the requested stage and replays any enabled stages between them. The report
+retains the original upstream metrics, methods, and output paths. A missing or
+modified registered output file identifies the stage to rerun. Changes to
+upstream inputs, compute settings, or analysis settings also invalidate the
+checkpoint. Studies saved with fingerprint version 5 or earlier need one fresh
+upstream run to create version 6 checkpoints.
+
+With `run.resume: true`, the integration benchmark can reuse a completed
+measurement of the same labels, batches and embedding coordinates. Annotation
+diagnostics and population identity run again when rebuilding the object so
+downstream analyses receive the cell-level evidence.
+
A whole-config validator rejects contradictory combinations — for example,
`compute.backend: auto` together with `compute.fallback_to_cpu: false`.
diff --git a/docs/reference/architecture.md b/docs/reference/architecture.md
index a556184e..356091fa 100644
--- a/docs/reference/architecture.md
+++ b/docs/reference/architecture.md
@@ -78,6 +78,12 @@ packages; a GPU method checks for a capable device — and skip with a reason wh
their backend is absent. This keeps method selection declarative (in YAML) and makes
backend availability a runtime concern rather than an install-time hard dependency.
+For a stage with a fixed default, set `default_method` on `DefaultMethodStage`.
+The planner and executor both use that registry name when a study omits the method.
+Stages with a custom selector can implement `_select_method_name` on `MethodDispatchStage`.
+Differential expression, differential abundance and enrichment share the primary
+contrast settings. Explicit method settings, including `paired: false`, take precedence.
+
R/Bioconductor methods share one abstraction, `cellquorum.methods.r_method.RAnalysisMethod`,
which centralizes Rscript resolution and R-package availability checks for the
edgeR, Milo, propeller, NicheNet, MultiNicheNet, and scDiagnostics adapters.
diff --git a/docs/reference/third-party-software.md b/docs/reference/third-party-software.md
index 86dbe616..1fe17a31 100644
--- a/docs/reference/third-party-software.md
+++ b/docs/reference/third-party-software.md
@@ -16,6 +16,7 @@ submit a paper.
|---|---|---|---|
| [anndata](https://github.com/scverse/anndata) | The data object every stage shares | BSD-3-Clause | Virshup et al., *JOSS* 2024, [10.21105/joss.04371](https://doi.org/10.21105/joss.04371) |
| [Scanpy](https://github.com/scverse/scanpy) | Neighbours, UMAP, scoring, plotting helpers | BSD-3-Clause | Wolf, Angerer & Theis, *Genome Biol* 2018, [10.1186/s13059-017-1382-0](https://doi.org/10.1186/s13059-017-1382-0) |
+| [umap-learn](https://github.com/lmcinnes/umap) | Approximate neighbor graphs and UMAP embeddings | BSD-3-Clause | McInnes et al., *JOSS* 2018, [10.21105/joss.00861](https://doi.org/10.21105/joss.00861) |
| [leidenalg](https://github.com/vtraag/leidenalg) | Leiden clustering | GPL-3.0-or-later | Traag, Waltman & van Eck, *Sci Rep* 2019, [10.1038/s41598-019-41695-z](https://doi.org/10.1038/s41598-019-41695-z) |
| [harmonypy](https://github.com/slowkow/harmonypy) | Harmony batch integration | GPL-3 (license file; the package metadata says MIT) | Korsunsky et al., *Nat Methods* 2019, [10.1038/s41592-019-0619-0](https://doi.org/10.1038/s41592-019-0619-0) |
| [kneed](https://github.com/arvkevi/kneed) | Choosing the number of principal components | BSD-3-Clause | Satopaa et al., *ICDCSW* 2011, [10.1109/ICDCSW.2011.20](https://doi.org/10.1109/ICDCSW.2011.20) |
diff --git a/envs/README.md b/envs/README.md
index 8d177dfd..bb48d055 100644
--- a/envs/README.md
+++ b/envs/README.md
@@ -5,62 +5,44 @@ CellQuorum uses a layered environment strategy to isolate incompatible dependenc
## Primary Environments
### `cellquorum-core`
-The main runtime environment containing CellQuorum's core dependencies. Install
-the CellQuorum wheel separately after creating the environment. Both CPU and GPU
-recipes bound AnnData to 0.12 for CellRank compatibility and use the same pinned
-upstream pygpcca source to avoid its released Jinja2 dependency conflict. PyTorch
-2.8.0's CPU wheel avoids downloading CUDA libraries for CPU workstations.
-Both recipes select Python 3.12/Linux x86_64 PyTorch wheels using direct URLs
-and SHA-256 hashes published in the official PyTorch wheel index. This avoids
-pip index-option lines that `conda-lock` cannot parse. These recipes are not
-native Windows, macOS, or ARM profiles.
-SciPy is pinned to 1.16.3 in both recipes, matching the validated runtime and
-preventing the mixed conda/pip solver from assigning it to both package managers
-at different versions.
-
-A fresh recipe-created CPU environment, followed by wheel/pytest/Cython
-installation, passed all nine installed-analysis, checkpoint, CellRank estimator,
-and PETSc signal-handling checks. `pip check` was clean. Cython is now included in
-the recipe itself. This is local Linux/WSL validation; it does not cover every
-optional method or establish a fully locked environment.
+The Linux x86_64 CPU and GPU controllers use Python 3.12, pandas 3.1.1,
+AnnData 0.13.4, SciPy 1.16.3 and CellRank 2.3.3. Both use a pinned pygpcca
+revision and a hashed PyTorch wheel. The CPU lock includes PyTorch 2.8.0+cpu
+and scib-metrics 0.5.10 for the full integration benchmark.
+
+Install the CPU lock and the CellQuorum wheel from the repository root.
+
+~~~bash
+conda-lock install --micromamba -n cellquorum-core envs/cellquorum-core.conda-lock.yml
+micromamba run -n cellquorum-core python -m pip install --no-deps dist/cellquorum-*.whl
+micromamba run -n cellquorum-core python -m pip check
+~~~
+
+The installed CPU workflow was checked on human pancreas integration and
+sample-level SCC signaling. The [reference results](../docs/benchmarks/pandas-anndata-review/README.md#cpu-workstation-validation)
+include the clean-run comparison and the checks on resumed results.
### `cellquorum-dev`
Development environment extending `cellquorum-core` with testing, linting, and build tools (pytest, ruff, mypy, sphinx, etc.).
### `cellquorum-gpu`
-GPU-enabled environment for GPU-accelerated analyses, extending `cellquorum-core` with CUDA-enabled PyTorch and scvi-tools.
-The recipe now selects CUDA 12.9, RAPIDS 26.4, CuPy 14.1.1, and the PyTorch
-2.8.0 CUDA-12.9 wheel. CellRank 2.3.2 uses AnnData 0.12 because its import fails
-with AnnData 0.13. The exact upstream pygpcca commit removes the obsolete Jinja2
-pin in its PyPI release. Git is included for that source installation.
+The Linux x86_64 analysis process runs pandas 3.1.1, AnnData 0.13.4 and the CUDA 12.9 PyTorch wheel. The RAPIDS worker uses cuDF 26.10.2, which requires pandas 3.0.x, and the same AnnData version. PCA, Harmony, neighbors, Leiden and UMAP exchange their inputs and results as indexed H5AD files. scVI and scANVI train in the analysis process. [Published cuDF requirements](https://pypi.org/pypi/cudf-cu12/26.10.2/json)
-A fresh Linux/WSL environment assembled with these corrections passed 14 tests:
-installed GPU analysis and checkpoint recovery, scVI/scANVI CUDA training, and
-CellRank/PETSc execution checks. Both GPU probes passed. This does not validate
-every optional method, hardware model, or production-scale dataset. A generated
-lockfile now records the resolved dependencies; fresh installation and workflow
-validation of that exact repository lock remain outstanding. Conda RAPIDS packages also expose Python
-metadata inconsistent with their CUDA-12/native conda dependencies, so `pip check`
-is not clean. That packaging issue remains unresolved.
-
-Create using flexible channel priority and the declared channels, then install
-the built CellQuorum wheel separately:
+From the checkout, with micromamba, conda-lock 4.0.5 and a CUDA 12.9-compatible driver installed, run:
```bash
-micromamba create --override-channels --channel-priority flexible -f envs/cellquorum-gpu.yml
-micromamba run -n cellquorum-gpu python -m pip install dist/cellquorum-*.whl
+conda-lock install --micromamba -n cellquorum-gpu envs/cellquorum-gpu.conda-lock.yml
+micromamba run -n cellquorum-gpu python -m pip install --no-deps dist/cellquorum-*.whl
+micromamba run -n cellquorum-gpu python -m venv "$HOME/.venvs/cellquorum-rapids"
+export CELLQUORUM_RAPIDS_PYTHON="$HOME/.venvs/cellquorum-rapids/bin/python"
+"$CELLQUORUM_RAPIDS_PYTHON" -m pip install -r envs/rapids-worker-requirements.txt
+"$CELLQUORUM_RAPIDS_PYTHON" -m pip check
+micromamba run -n cellquorum-gpu python -m pip check
micromamba run -n cellquorum-gpu python scripts/verify_gpu.py --profile all
```
-Activate the environment or use `micromamba run`; direct interpreter invocation
-without activation initially failed CUDA header discovery during validation.
-
-Validate both accelerator families with `python scripts/verify_gpu.py --profile all`,
-then run installed-wheel acceptance with `CELLQUORUM_ACCEPTANCE_BACKEND=gpu`.
-The probes exercise device operations; the acceptance test exercises PCA,
-clustering, saved outputs, and checkpoint continuation. Neither establishes
-production-scale memory limits or validates every optional method.
+Export the same worker path in shells that launch the pipeline. The GPU Docker target sets it in the image. Environments with compatible in-process RAPIDS can still use the native backend. Set `compute.backend: gpu` and `compute.fallback_to_cpu: false` to require GPU acceleration for the routed methods. The [human pancreas reference run](../docs/benchmarks/pandas-anndata-review/README.md#gpu-workstation-validation) exercises the installed wheel and both CUDA stacks.
### `cellquorum-r`
R bridge environment for the R/Bioconductor methods dispatched over the Rscript
@@ -175,8 +157,7 @@ make lock LOCK_PYTHON=python
conda-lock 4.0.5, including 2.27 used by JAX and NVIDIA wheels. It retains the
tool's glibc ceiling and target-version filtering. Changing conda-lock versions
requires reviewing this workaround; the wrapper refuses an unreviewed version.
-The repository contains Linux CPU (417 package records) and GPU (528 package
-records) locks. CI checks package ownership across conda/pip, artifact hashes,
+The repository contains Linux CPU and GPU analysis-environment locks. CI checks package ownership across conda/pip, artifact hashes,
immutable Git revisions, direct pip requirements, and CPU/GPU profile selection.
These structural checks do not establish successful installation, complete recipe
freshness, runtime compatibility, or scientific accuracy.
@@ -188,11 +169,7 @@ conda-lock install --micromamba -n cellquorum-core-locked envs/cellquorum-core.c
conda-lock install --micromamba -n cellquorum-gpu-locked envs/cellquorum-gpu.conda-lock.yml
```
-The previous temporary validation environments and unfinished test logs are no
-longer available. Fresh installation and installed-wheel acceptance of these
-exact repository locks must therefore be repeated before calling them validated
-deployment profiles. The earlier recipe-based test results above remain separate
-evidence.
+The GPU instructions above include the separate RAPIDS worker.
Locks are solved for **linux-64 only** — the platform the Docker image builds
on. This is deliberate: the GPU env selects Linux CUDA/RAPIDS packages and a
@@ -241,3 +218,5 @@ SampleQC from assessment-only to automatic cell removal.
This pins SampleQC's source, not its full transitive R environment. A freshly
provisioned and locked R profile remains a deployment requirement.
+
+The [extended pancreas installation guide](../docs/benchmarks/reference-studies.md#extended-pancreas-lineage-from-a-clean-python-312-install) covers the lineage extra, the isolated CytoTRACE 2 worker and model weights.
diff --git a/envs/cellquorum-core.conda-lock.yml b/envs/cellquorum-core.conda-lock.yml
index b21e25d4..f91b660c 100644
--- a/envs/cellquorum-core.conda-lock.yml
+++ b/envs/cellquorum-core.conda-lock.yml
@@ -13,7 +13,7 @@
version: 1
metadata:
content_hash:
- linux-64: 48b7764bb4849815c206bcba3897b6587f0d1de96d0771cba7ac8535104fe3ef
+ linux-64: 811bd1a4fcf34a22281f90647d41398d68077a532d2690a71408891f8f938a7e
channels:
- url: conda-forge
used_env_vars: []
@@ -78,37 +78,40 @@ package:
category: main
optional: false
- name: anndata
- version: 0.12.19
+ version: 0.13.4
manager: conda
platform: linux-64
dependencies:
array-api-compat: '>=1.7.1'
- h5py: '>=3.8'
+ h5py: '>=3.11'
legacy-api-wrap: ''
natsort: ''
- numpy: '>=1.26'
+ numpy: '>=2.1'
packaging: '>=24.2'
- pandas: '>=2.1.0,!=2.1.2,<3'
- python: '>=3.11'
- scipy: '>=1.12,!=1.17.0'
- scverse-misc: '>=0.0.3'
- zarr: '>=2.18.7,!=3.0.*'
- url: https://conda.anaconda.org/conda-forge/noarch/anndata-0.12.19-pyhd8ed1ab_0.conda
+ pandas: '>=2.3'
+ pydantic-settings: ''
+ python: '>=3.12'
+ python-dotenv: ''
+ scipy: '>=1.14,!=1.17.0'
+ scverse-misc: '>=0.1.0'
+ typing_extensions: '>=4.16'
+ zarr: '>=3.1'
+ url: https://conda.anaconda.org/conda-forge/noarch/anndata-0.13.4-pyhd8ed1ab_0.conda
hash:
- md5: cb6497a1f1f2e3f83256e8eadb995bd2
- sha256: f9253f582480d2f49e8b1ca49922ccdaa1815674bd2b83742c11975f1116a001
+ md5: 8c7b633e969cec15952d4318d7cbb66f
+ sha256: c325ebd706c72e7db772f5311f67373cbc2b1ad35b0e3fe0c51f803b8e8a41d6
category: main
optional: false
- name: annotated-doc
- version: 0.0.5
+ version: 0.0.6
manager: conda
platform: linux-64
dependencies:
- python: '>=3.10'
- url: https://conda.anaconda.org/conda-forge/noarch/annotated-doc-0.0.5-pyhcf101f3_0.conda
+ python: '>=3.11'
+ url: https://conda.anaconda.org/conda-forge/noarch/annotated-doc-0.0.6-pyh5ded981_0.conda
hash:
- md5: d68ec17cb915cab4642357417ec0a104
- sha256: a0a320c709fded9b5178108dedebf81a1f5a5d9c7720e3af50c1ab058dadd296
+ md5: 9fca85fe0c3f76ab2199f576d7d115c0
+ sha256: ade8040829ed573c628afb85868a74bd03df7f3c663be318f6e5bbad2e905f78
category: main
optional: false
- name: annotated-types
@@ -866,7 +869,7 @@ package:
category: main
optional: false
- name: decoupler-py
- version: 2.3.1
+ version: 2.2.0
manager: conda
platform: linux-64
dependencies:
@@ -875,15 +878,15 @@ package:
docrep: ''
ipywidgets: ''
marsilea: ''
- python: '>=3.12'
+ python: '>=3.11'
python-igraph: ''
requests: ''
scanpy: ''
xgboost: ''
- url: https://conda.anaconda.org/conda-forge/noarch/decoupler-py-2.3.1-pyhd8ed1ab_0.conda
+ url: https://conda.anaconda.org/conda-forge/noarch/decoupler-py-2.2.0-pyhd8ed1ab_0.conda
hash:
- md5: 5f873c51e540df493e9c21c61dea146c
- sha256: ba9801b7a32aac3e8661e7f68a27bbfc28e7cb57a092a8992b2ee84ee444304d
+ md5: 47fb2cb7ca1f606ca4e5bc98a96ba597
+ sha256: f20d633d2d42e0a9dbf8d5ec18a70e234d046aba4ea3f68b540ad3756e7dc27e
category: main
optional: false
- name: docrep
@@ -1039,10 +1042,10 @@ package:
manager: conda
platform: linux-64
dependencies: {}
- url: https://conda.anaconda.org/conda-forge/noarch/font-ttf-ubuntu-0.83-h77eed37_3.conda
+ url: https://conda.anaconda.org/conda-forge/noarch/font-ttf-ubuntu-0.83-h77eed37_4.conda
hash:
- md5: 49023d73832ef61042f6a237cb2687e7
- sha256: 2821ec1dc454bd8b9a31d0ed22a7ce22422c0aef163c59f49dfdf915d0f0ca14
+ md5: 1df978a6efeb98fa80b96f2860b9b280
+ sha256: a7384ac883b1acf492fd1aef27431f0f1a647bbb8b5709a1bfab11b0a765f5e5
category: main
optional: false
- name: fontconfig
@@ -1795,6 +1798,7 @@ package:
azure-core-cpp: '>=1.16.4,<1.16.5.0a0'
azure-identity-cpp: '>=1.13.3,<1.13.4.0a0'
azure-storage-blobs-cpp: '>=12.19.0,<12.19.1.0a0'
+ azure-storage-common-cpp: '>=12.15.0,<12.15.1.0a0'
azure-storage-files-datalake-cpp: '>=12.17.0,<12.17.1.0a0'
bzip2: '>=1.0.8,<2.0a0'
glog: '>=0.7.1,<0.8.0a0'
@@ -1811,13 +1815,13 @@ package:
lz4-c: '>=1.10.0,<1.11.0a0'
orc: '>=2.3.1,<2.3.2.0a0'
simdjson: '>=5.1.1,<5.2.0a0'
- snappy: '>=1.2.2,<1.3.0a0'
+ snappy: '>=1.3.1,<1.4.0a0'
uriparser: '>=0.9.8,<1.0a0'
zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-26.0.1-hb81f139_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-26.0.1-h4f621f8_5_cpu.conda
hash:
- md5: 5e467a924e4e3a48ed03fda0f3650e88
- sha256: 76638e36735c2c8f968b090814c1b583e2c8f2b56b05e155c4b82fb9f286e652
+ md5: ecde9278c41bee3cde9ddc99ee7f8d0e
+ sha256: ce351947c34f04e4b16f0d98441e02793471e043f630670d6258cb18e60a4bf5
category: main
optional: false
- name: libarrow-acero
@@ -1830,10 +1834,10 @@ package:
libarrow-compute: 26.0.1
libgcc: '>=15'
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-acero-26.0.1-h35cad7b_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-acero-26.0.1-h35cad7b_5_cpu.conda
hash:
- md5: a3e52de423eacd0f8fc45411b96e730c
- sha256: 3e895c0ddfd462b691d1d21fcd045cd59c5e34c3a348fb8c4eba05cebf27c037
+ md5: 74b436e99ef9bfda9c99592f586acf89
+ sha256: 339c364cd77e7afa72dbbcb9dc906a25c64421889bb2c334dc1c5016110bdbfd
category: main
optional: false
- name: libarrow-compute
@@ -1846,12 +1850,12 @@ package:
libgcc: '>=15'
libre2-11: '>=2025.11.5'
libstdcxx: '>=15'
- libutf8proc: '>=2.11.3,<2.12.0a0'
+ libutf8proc: '>=2.12.0,<2.13.0a0'
re2: ''
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-compute-26.0.1-h2e1e52a_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-compute-26.0.1-hc9b89cc_5_cpu.conda
hash:
- md5: f01eb652e0b2de6156317edd2fbd9d39
- sha256: 0d99a653576daad68e2bc356719ec2ea708e143c86f6b8a4fd9a9c7c4bacf9af
+ md5: 22def3abed674d30d38edfad0416f289
+ sha256: a1ca1f213dee003cee364b69c9ad596e3e01e33d87051c2be0626bd59a426f62
category: main
optional: false
- name: libarrow-dataset
@@ -1866,10 +1870,10 @@ package:
libgcc: '>=15'
libparquet: 26.0.1
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-dataset-26.0.1-h35cad7b_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-dataset-26.0.1-h35cad7b_5_cpu.conda
hash:
- md5: bf611bb0386926bc33641ac13c3b371b
- sha256: a57b26262304eb4afe3c11821f4391885a42dee58da8f733a44b058cb7c6e0fa
+ md5: 74ed7c8f31879eba79eaf6884c97aea2
+ sha256: 7d3c12fd832d10ea2f0e81461f0063420bbbf66219213eec97d227a6eec7fb6f
category: main
optional: false
- name: libarrow-s3
@@ -1883,10 +1887,10 @@ package:
libarrow: 26.0.1
libgcc: '>=15'
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-s3-26.0.1-hb32f990_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-s3-26.0.1-hb32f990_5_cpu.conda
hash:
- md5: 9d862a322d87f13aa11ce5191192c644
- sha256: 2e7a8fd5e018e8a450d988da69821759c4876ba809ec8a7ecd15a5167ddd1835
+ md5: b638106455b70e550b443bb532804902
+ sha256: 509cdde99f5b0ba5f7bd76ae7f151021ab5957ca56c39ea265db42b818dfb8d9
category: main
optional: false
- name: libarrow-substrait
@@ -1902,10 +1906,10 @@ package:
libgcc: '>=15'
libprotobuf: '>=7.35.1,<7.35.2.0a0'
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-substrait-26.0.1-hbad6200_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-substrait-26.0.1-hbad6200_5_cpu.conda
hash:
- md5: 649c9bda1221f60d347a9f5338fcd190
- sha256: d32638cb0700029202ca6e2563499a823707912add0c42d68b563bc821f94c03
+ md5: 4be0c900d1f2acb36d2954c74ec4ff02
+ sha256: 18764bdb42e52ae903e431b0f40f27ee562324174a3f2bc16ef3c0658dfba32c
category: main
optional: false
- name: libblas
@@ -2051,42 +2055,42 @@ package:
category: main
optional: false
- name: libclang-cpp23.1
- version: 23.1.4
+ version: 23.1.5
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
libgcc: '>=15'
- libllvm23: '>=23.1.4,<23.2.0a0'
+ libllvm23: '>=23.1.5,<23.2.0a0'
libstdcxx: '>=15'
libxml2: ''
libxml2-16: '>=2.15.4'
libzlib: '>=1.3.2,<2.0a0'
zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libclang-cpp23.1-23.1.4-default_h0acdd01_1.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libclang-cpp23.1-23.1.5-default_h0acdd01_0.conda
hash:
- md5: c3831139a4d948c287b4336f1b2059fd
- sha256: 93dde2595e2c339d687be470e1f458f7641a52bda06762055db19237cbf40fac
+ md5: 14c1b208c11eea04cef332636b002394
+ sha256: 0c506f0e598c90e6fe374188e6e102a1346dcbf2b43d25e3db3e3975700d8a5d
category: main
optional: false
- name: libclang13
- version: 23.1.4
+ version: 23.1.5
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libclang-cpp23.1: ==23.1.4
+ libclang-cpp23.1: ==23.1.5
libgcc: '>=15'
- libllvm23: '>=23.1.4,<23.2.0a0'
+ libllvm23: '>=23.1.5,<23.2.0a0'
libstdcxx: '>=15'
libxml2: ''
libxml2-16: '>=2.15.4'
libzlib: '>=1.3.2,<2.0a0'
zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libclang13-23.1.4-default_h6e6c352_1.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libclang13-23.1.5-default_h782a7a8_0.conda
hash:
- md5: 2ed7cfa18d17f82b546ee2ed46bf2f7b
- sha256: 0d973a298e2e9efb1f563e0e1f2a39cc7704c4393e1aad1128c95b9b85bff1d1
+ md5: 77a10c771b02cfd2f433d43b64298892
+ sha256: 727fb7f72d8f9aedea7a00e0d043bbc87153c9aa6951dd21e446e75378053776
category: main
optional: false
- name: libcolamd
@@ -2153,16 +2157,16 @@ package:
category: main
optional: false
- name: libdeflate
- version: '1.25'
+ version: '1.26'
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libgcc: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libdeflate-1.25-hd45a770_1.conda
+ libgcc: '>=15'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libdeflate-1.26-h351257e_0.conda
hash:
- md5: 40f9b31aa9cf007789867df0decd0492
- sha256: 82e134c8a08b1eed9a2ed8ab578b89aa1730dcde3dea8dd87645ed0637878e54
+ md5: d0c17fc12939a990f0865b7a14b49cfe
+ sha256: f437f68abb1fa1dededd082a0c2800bec40495e01cf9ec46f4e693e34d60b644
category: main
optional: false
- name: libdrm
@@ -2235,17 +2239,17 @@ package:
category: main
optional: false
- name: libevent
- version: 2.1.12
+ version: 2.1.13
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
libgcc: '>=15'
- openssl: '>=3.5.8,<4.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libevent-2.1.12-h5348a74_2.conda
+ openssl: '>=3.5.9,<4.0a0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libevent-2.1.13-h5348a74_0.conda
hash:
- md5: 2e699c6e151045b2cfea34b53ed69439
- sha256: f874779772878f3b509c607de9089c282f71232fde4cac876fc660d29670f28d
+ md5: 53b60cc0d314d0f1ebfdbbf7d8f6e313
+ sha256: 57ed3320311043c9bc829825fa6e30db3e17307acdd8f5b7bbc917bb36247bf4
category: main
optional: false
- name: libexpat
@@ -2407,7 +2411,7 @@ package:
category: main
optional: false
- name: libglib
- version: 2.90.0
+ version: 2.90.1
manager: conda
platform: linux-64
dependencies:
@@ -2417,10 +2421,10 @@ package:
libiconv: '>=1.18,<2.0a0'
libzlib: '>=1.3.2,<2.0a0'
pcre2: '>=10.47,<10.48.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libglib-2.90.0-h569388d_0.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libglib-2.90.1-h569388d_1.conda
hash:
- md5: c85e9539551f1fd21104e518d940c244
- sha256: 9e688e046550415c030c3d08ed9a3fd13a51d7bc56295283b858b6c3070bb9f8
+ md5: 2b4731a8f8cd069c1d177f972e9d0d99
+ sha256: 463a24904fd83a3223c9f292211975bba405838ff32e6179df04a3cca202a2b3
category: main
optional: false
- name: libglvnd
@@ -2818,10 +2822,10 @@ package:
libthrift: '>=0.25.0,<0.25.1.0a0'
openssl: '>=3.5.9,<4.0a0'
simdjson: '>=5.1.1,<5.2.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libparquet-26.0.1-h2d6894f_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libparquet-26.0.1-h2d6894f_5_cpu.conda
hash:
- md5: d323682fdf093fec93a918e1c3b4d7a8
- sha256: 17b77aa533b3496c075b2428743c41a3e1baee2485fd6f85583402664e5d2ff8
+ md5: fe532df58c640bd0eb6215bc45a3e4ce
+ sha256: 96d1d46a7b7d4df43a49b8555195965b8587075df1b5d88adc7b7bcc326fe4a8
category: main
optional: false
- name: libpciaccess
@@ -2862,11 +2866,11 @@ package:
libgcc: '>=15'
libstdcxx: '>=15'
openldap: '>=2.6.13,<2.7.0a0'
- openssl: '>=3.5.8,<4.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libpq-18.6-h3883810_3.conda
+ openssl: '>=3.5.9,<4.0a0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libpq-18.6-h3883810_4.conda
hash:
- md5: cc998047391971a452fed680158329f3
- sha256: 60fa3b973f6f68a367171336382d801404ee7b38c987c3b6a55c207086f736de
+ md5: 13266bfcec35113306ccb14fb578dfae
+ sha256: 94b9a4abf3bd67bb772d8050ae3933b7987b0a9a28ca354b35c45c2ebd93b796
category: main
optional: false
- name: libprotobuf
@@ -2952,10 +2956,10 @@ package:
libabseil: '*,>=20260817.0,<20260818.0a0'
libgcc: '>=15'
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/libre2-11-2025.11.05-h02ca46f_3.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libre2-11-2025.11.05-h02ca46f_4.conda
hash:
- md5: ad950bb7d6c2b0a2cb62d9af66c9a082
- sha256: 8310dbf66d801a577349eb6268a6ef83accc5f5a042b2911a7d7273a9f991105
+ md5: 2992bcf0ff620215b8427fcafad4afe8
+ sha256: 0c49508a9525f89e9195155e106ae8c9fb57ddf46e5111f98d7f3227976c6a15
category: main
optional: false
- name: libscotch
@@ -3097,15 +3101,15 @@ package:
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libevent: '>=2.1.12,<2.1.13.0a0'
+ libevent: '>=2.1.13,<2.1.14.0a0'
libgcc: '>=15'
libstdcxx: '>=15'
libzlib: '>=1.3.2,<2.0a0'
openssl: '>=3.5.9,<4.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libthrift-0.25.0-h22fcfe2_1.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libthrift-0.25.0-h06f262d_2.conda
hash:
- md5: 623422483280b48e40139e11f7991f31
- sha256: 70c6e9d07c960739ebbac5519225aa0f80556ad1755597c99f7188eeeb37fa33
+ md5: 3a01b39543e4b3e1e8dcb74f0935727a
+ sha256: e3ea13e90078bd3a87162a2b4c030507f0d0bdad1a6064f9ab5e19321b944f45
category: main
optional: false
- name: libtiff
@@ -3114,19 +3118,19 @@ package:
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- lerc: '>=4.2.0,<5.0a0'
- libdeflate: '>=1.25,<1.26.0a0'
+ lerc: '>=4.3.0,<5.0a0'
+ libdeflate: '>=1.26,<1.27.0a0'
libgcc: '>=15'
- libjpeg-turbo: '>=3.2.0,<4.0a0'
+ libjpeg-turbo: '>=3.2.1,<4.0a0'
liblzma: '>=5.8.3,<6.0a0'
libstdcxx: '>=15'
libwebp-base: '>=1.6.0,<2.0a0'
libzlib: '>=1.3.2,<2.0a0'
zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libtiff-4.7.2-hcc2c06a_1.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libtiff-4.7.2-h6d37cbf_2.conda
hash:
- md5: 0907d876f460ebc941ae590338b6102f
- sha256: 10e125da82ca93e09191e66e5a94d566b2cf8d4024e2a0db3a9114297447e0d9
+ md5: 18023fe8fbec4f031ca88444572590e4
+ sha256: 3406eed96ee929309ecd3937308d5a81302f4ca2eddd30e832071cbe9121a6c7
category: main
optional: false
- name: libudev1
@@ -3161,16 +3165,16 @@ package:
category: main
optional: false
- name: libutf8proc
- version: 2.11.3
+ version: 2.12.0
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libgcc: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libutf8proc-2.11.3-hfe17d71_0.conda
+ libgcc: '>=15'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libutf8proc-2.12.0-hc4fb7df_0.conda
hash:
- md5: 1247168fe4a0b8912e3336bccdbf98a5
- sha256: ecbf4b7520296ed580498dc66a72508b8a79da5126e1d6dc650a7087171288f9
+ md5: a7d99b24a07fda1a6442ad64b759bea5
+ sha256: 0c86e7b7fd600b05ac84cf1bc049e3a562e3bdd5c37433386a611060b7856fb9
category: main
optional: false
- name: libuuid
@@ -3225,10 +3229,10 @@ package:
pthread-stubs: ''
xorg-libxau: '>=1.0.12,<2.0a0'
xorg-libxdmcp: '>=1.1.5,<2.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libxcb-1.17.0-hb83e432_2.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libxcb-1.17.0-hb83e432_3.conda
hash:
- md5: 61f0ffba9161cbb3a77722869b21e4bb
- sha256: 7b49e6fd2a584b7f072943e6adf4149677af5121246975278804782d37752837
+ md5: 661d9e1392a1b672d19d76df3a1feb07
+ sha256: 38021f35547021762c04860d1c9f2d3186b83e090e5f1773216045de06cf10cc
category: main
optional: false
- name: libxcrypt
@@ -3778,22 +3782,22 @@ package:
category: main
optional: false
- name: numpy
- version: 2.5.4
+ version: 2.3.5
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
libblas: '>=3.9.0,<4.0a0'
libcblas: '>=3.9.0,<4.0a0'
- libgcc: '>=15'
+ libgcc: '>=14'
liblapack: '>=3.9.0,<4.0a0'
- libstdcxx: '>=15'
+ libstdcxx: '>=14'
python: ''
python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/numpy-2.5.4-py312he827f4e_0.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/numpy-2.3.5-py312h33ff503_1.conda
hash:
- md5: 759b701f4b548ff6d06794d99788e6bb
- sha256: 8d562d8fcc44a93c65793a361f39a75712bb3cf2aa1d1a09bcc8ef072e54a1c3
+ md5: 03baecffb72fa96fe234fd505908065f
+ sha256: 76ad6a6f4761084b074a587fe1512956891f04b5250cec0fd39aca0f39ad122b
category: main
optional: false
- name: omegaconf
@@ -3886,13 +3890,13 @@ package:
libstdcxx: '>=15'
libzlib: '>=1.3.2,<2.0a0'
lz4-c: '>=1.10.0,<1.11.0a0'
- snappy: '>=1.2.2,<1.3.0a0'
+ snappy: '>=1.3.1,<1.4.0a0'
tzdata: ''
zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/orc-2.3.1-hdc56c17_1.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/orc-2.3.1-h7b2d648_2.conda
hash:
- md5: d65dbc4db24e05c2f1c9664fb34e23d1
- sha256: 7e0edd79df3c6c70fa46e4322cd92c107cd617e3b3afc8d809b561ca105ad77d
+ md5: 82f0a2f57530e0bbfbc83aa46e1d851a
+ sha256: fb29474cac4c60dbc929e539c391303b0f47fb846459e36edc44ae841b9858f3
category: main
optional: false
- name: packaging
@@ -3908,23 +3912,21 @@ package:
category: main
optional: false
- name: pandas
- version: 2.3.3
+ version: 3.1.1
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- numpy: '>=1.22.4,>=1.23,<3'
- python: '>=3.12,<3.13.0a0'
- python-dateutil: '>=2.8.2'
- python-tzdata: '>=2022.7'
+ libgcc: '>=15'
+ libstdcxx: '>=15'
+ numpy: '>=1.25,<3,>=2.0.2'
+ python: ''
+ python-dateutil: '>=2.9.0'
python_abi: 3.12.*
- pytz: '>=2020.1'
- url: https://conda.anaconda.org/conda-forge/linux-64/pandas-2.3.3-py312hf79963d_1.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/pandas-3.1.1-np2py312h91ec553_0.conda
hash:
- md5: e597b3e812d9613f659b7d87ad252d18
- sha256: f633d5f9b28e4a8f66a6ec9c89ef1b6743b880b0511330184b4ab9b7e2dda247
+ md5: d3c0f076ef954125c9c3db84794bc51b
+ sha256: bcbf9b1c9c040b343791536fdd6c5acd87e73f0c2d58859c9164756528626b34
category: main
optional: false
- name: parmetis
@@ -4134,6 +4136,21 @@ package:
sha256: 8bbb25f131001457e5fd19385000e9d6a8047223a35f4c507d92113eb68d42f9
category: main
optional: false
+- name: pooch
+ version: 1.9.0
+ manager: conda
+ platform: linux-64
+ dependencies:
+ packaging: '>=20.0'
+ platformdirs: '>=2.5.0'
+ python: '>=3.10'
+ requests: '>=2.19.0'
+ url: https://conda.anaconda.org/conda-forge/noarch/pooch-1.9.0-pyhd8ed1ab_0.conda
+ hash:
+ md5: dd4b6337bf8886855db6905b336db3c8
+ sha256: 081e52c4612830bf1fd4a9c78eebaf335d1385d74ddfd328b1b2f26b983848eb
+ category: main
+ optional: false
- name: prometheus-cpp
version: 1.3.0
manager: conda
@@ -4537,18 +4554,6 @@ package:
sha256: ee9c2922e07afc85fc82d5fa82c9ac2c79da3be3283a5e17bf2f2ae38dbd02fb
category: main
optional: false
-- name: pytz
- version: '2026.6'
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.11'
- url: https://conda.anaconda.org/conda-forge/noarch/pytz-2026.6-pyh5ded981_0.conda
- hash:
- md5: 9b60b8aadafa78f7581c49a057764e7f
- sha256: 624d4243a9a2ebec6349587d34bd64c8b04ab6ed174acf102e02a441afde5ae0
- category: main
- optional: false
- name: pyyaml
version: 6.0.3
manager: conda
@@ -4668,10 +4673,10 @@ package:
platform: linux-64
dependencies:
libre2-11: 2025.11.05
- url: https://conda.anaconda.org/conda-forge/linux-64/re2-2025.11.05-hbf548d4_3.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/re2-2025.11.05-hbf548d4_4.conda
hash:
- md5: eac6d593030ba95beda012452003ed53
- sha256: a77bfa9d62f1df59d46e8de8961ebc6b618ea055ccd9e3f050ace4d10f459ff8
+ md5: f05fcc2734b47231b9a5623c6d4f43ac
+ sha256: 54a40d376a64b91b1a91decb1d4986f189cfb26fd4cdaddd850e5aafa092500d
category: main
optional: false
- name: readline
@@ -4834,24 +4839,6 @@ package:
sha256: 90861e7f903ad48a25a8f5d0d4a74b010b517c47026ca32fe385e833542a1055
category: main
optional: false
-- name: scikit-misc
- version: 0.5.3
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=15'
- libgfortran: ''
- libgfortran5: '>=15.3.0'
- numpy: '>=1.25,<3'
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/scikit-misc-0.5.3-py312hc3447ac_0.conda
- hash:
- md5: 5a3d17773efbf1286b9a4732c9bae445
- sha256: 3d554ffb3eba5c38f2e89ac325dc03362c1b6f24ecd509d4869c4758f2012916
- category: main
- optional: false
- name: scipy
version: 1.16.3
manager: conda
@@ -5042,17 +5029,17 @@ package:
category: main
optional: false
- name: snappy
- version: 1.2.2
+ version: 1.3.1
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
libgcc: '>=15'
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/snappy-1.2.2-h34e00bb_2.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/snappy-1.3.1-h34e00bb_0.conda
hash:
- md5: 2b7e31ffbd476eecf33b3b833692339c
- sha256: 8635d10c7f127808b09c0c8e77b21a255829e0934e49326548c949f7e0551f09
+ md5: 9fe09f49cf48a5e73ef0c8dbfad570ba
+ sha256: a724d5312f1c8fe84be83b176b20fa171b3106cdfec1648cc3f32888397ecb09
category: main
optional: false
- name: stack_data
@@ -5158,28 +5145,28 @@ package:
category: main
optional: false
- name: threadpoolctl
- version: 3.7.0
+ version: 3.8.0
manager: conda
platform: linux-64
dependencies:
python: '>=3.11'
- url: https://conda.anaconda.org/conda-forge/noarch/threadpoolctl-3.7.0-pyhc455866_0.conda
+ url: https://conda.anaconda.org/conda-forge/noarch/threadpoolctl-3.8.0-pyhc455866_0.conda
hash:
- md5: 8c05910e780502e362d658f00f5ec10f
- sha256: a32d17fe979e59fe1295875b5d4f0431a76dfbbfcee5b01e657a8aa3c4a1b5d1
+ md5: a84c9b62a40f5d6f1c07df8117ae770b
+ sha256: 6dde58ec2e194419a7381622a3177fa9d4c424823e9c65b04b905892ce7d7779
category: main
optional: false
- name: tifffile
- version: 2026.10.15
+ version: 2026.11.8
manager: conda
platform: linux-64
dependencies:
numpy: '>=2.1'
python: '>=3.12'
- url: https://conda.anaconda.org/conda-forge/noarch/tifffile-2026.10.15-pyhc364b38_0.conda
+ url: https://conda.anaconda.org/conda-forge/noarch/tifffile-2026.11.8-pyhc364b38_0.conda
hash:
- md5: 69311a22cf97c3c9840108555b83e4f7
- sha256: 74205e597f33d5aa55dc535481cff7d7341b8026c1698be0d0c62f1e7c42110f
+ md5: da6fc49f7ac67a900f93a51ddf15d585
+ sha256: d5505801ed156186a416ac650406c55951e1fc7bb8a31c5c7442d65c415a3e99
category: main
optional: false
- name: tk
@@ -5237,19 +5224,20 @@ package:
category: main
optional: false
- name: typer
- version: 0.27.3
+ version: 0.29.0
manager: conda
platform: linux-64
dependencies:
annotated-doc: '>=0.0.2'
colorama: ''
+ pydantic: '>=2.5.3'
python: '>=3.11'
rich: '>=13.8.0'
shellingham: '>=1.3.0'
- url: https://conda.anaconda.org/conda-forge/noarch/typer-0.27.3-pyh5ded981_0.conda
+ url: https://conda.anaconda.org/conda-forge/noarch/typer-0.29.0-pyh5ded981_0.conda
hash:
- md5: 3251bd81ef66ec6432b3159944c70d57
- sha256: 50cef25b46782fbb8291427459ed60c701165b0b70bdb3aebae2503e9915f66a
+ md5: 50b1de5cb639dff936c40b1577015f9a
+ sha256: e0e98c372dc47488d8c49c47800c8ce8da29eb6a66ebd4befd03c7b6fb71f44d
category: main
optional: false
- name: typing-extensions
@@ -5775,13 +5763,13 @@ package:
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libgcc: '>=14'
- xorg-libx11: '>=1.8.12,<2.0a0'
- xorg-libxext: '>=1.3.6,<2.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/xorg-libxxf86vm-1.1.7-hb03c661_0.conda
+ libgcc: '>=15'
+ xorg-libx11: '>=1.8.14,<2.0a0'
+ xorg-libxext: '>=1.3.7,<2.0a0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/xorg-libxxf86vm-1.1.7-h7cc23a3_1.conda
hash:
- md5: 665d152b9c6e78da404086088077c844
- sha256: 64db17baaf36fa03ed8fae105e2e671a7383e22df4077486646f7dbf12842c9f
+ md5: ac05771f924d124349ef9cc3140e9876
+ sha256: 264ddaeba8de4fe17fb25d10c6f6c2aac5d8d41d8883a0976dae59b528ef9520
category: main
optional: false
- name: xorg-xorgproto
@@ -5903,13 +5891,14 @@ package:
category: main
optional: false
- name: cell2cell
- version: 0.8.4
+ version: 0.9.0
manager: pip
platform: linux-64
dependencies:
gseapy: '>=1.0.3'
kneed: '*'
matplotlib: '>=3.2.0'
+ natsort: '*'
networkx: '>=2.3'
numpy: '>=1.16'
openpyxl: '>=2.6.2'
@@ -5923,13 +5912,13 @@ package:
tqdm: '*'
umap-learn: '*'
xlrd: '>=1.1'
- url: https://files.pythonhosted.org/packages/39/40/d0c5b144ff8f880e4e03da35969ce6ff055cde3e6544b722eda24f19c357/cell2cell-0.8.4-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/6f/70/7b36729239f848ea264c2493c52e7247d8eaf8f6e0582767125a395ea40a/cell2cell-0.9.0-py3-none-any.whl
hash:
- sha256: 506dcffadafc06d2dc3358213e8df5ca16ab5cca1c8f48c03640a76f341d56ca
+ sha256: 4a2d2e3cfb7706def00183fe8205ae08a1106bf166b33ea6c23d9b842098f4d6
category: main
optional: false
- name: cellrank
- version: 2.3.2
+ version: 2.3.3
manager: pip
platform: linux-64
dependencies:
@@ -5950,9 +5939,9 @@ package:
seaborn: '>=0.13'
session-info2: '>=0.2'
wrapt: '>=1.16'
- url: https://files.pythonhosted.org/packages/63/02/3f07b87bdcc0354f7b4bef973b2ecc60676035ec04754c4cc7457ff9a4c1/cellrank-2.3.2-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/71/9d/45fa014e466d80ca492f4a839b3b416343c6ddd9e08fb03e859fc3e04697/cellrank-2.3.3-py3-none-any.whl
hash:
- sha256: 7a29e03cabac4a7bc80335e48341a10f69383d36283ff84dc8e0e9a086932cf5
+ sha256: 351578a18beb36bbba24e1b79cb0545800ea34e0c960497d7878ae8f297a1f8b
category: main
optional: false
- name: celltypist
@@ -5973,14 +5962,31 @@ package:
sha256: 3953f96a4476ed1dd2807446d97d3945e8905bbdac6bfef0d20b84a9a00d1101
category: main
optional: false
+- name: chex
+ version: 0.1.90
+ manager: pip
+ platform: linux-64
+ dependencies:
+ absl-py: '>=0.9.0'
+ jax: '>=0.4.27'
+ jaxlib: '>=0.4.27'
+ numpy: '>=1.24.1'
+ setuptools: '*'
+ toolz: '>=0.9.0'
+ typing-extensions: '>=4.2.0'
+ url: https://files.pythonhosted.org/packages/6f/3d/46bb04776c465cea2dd8aa2d4b61ab610b707f798f47838ef7e6105b025c/chex-0.1.90-py3-none-any.whl
+ hash:
+ sha256: fce3de82588f72d4796e545e574a433aa29229cbdcf792555e41bead24b704ae
+ category: main
+ optional: false
- name: click
- version: 8.5.0
+ version: 8.5.1
manager: pip
platform: linux-64
dependencies: {}
- url: https://files.pythonhosted.org/packages/58/50/6c0d534c5f134586a8e1ba4e330569e32f057e33372ae556463212fb4cd3/click-8.5.0-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/31/7d/164b7ffc8624fe9efb4221bf9af485d5c1e41b9ae118045d2b6dc5457d8b/click-8.5.1-py3-none-any.whl
hash:
- sha256: 255bc9599cf7748b4b1a446ccc735421bd08a2ae529a8b88597d3de5664ee360
+ sha256: 8bb5376d2006950386f8c6427613334a2acadc5f559cb80e638d00b4df957e37
category: main
optional: false
- name: deprecated
@@ -5995,13 +6001,13 @@ package:
category: main
optional: false
- name: filelock
- version: 4.2.9
+ version: 4.2.10
manager: pip
platform: linux-64
dependencies: {}
- url: https://files.pythonhosted.org/packages/00/05/76184fe08339723d643ab47887a8e4de8f2921ca4e008617ab7490af7857/filelock-4.2.9-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/ef/90/5d0096a325aa0f83ba7a3968bfb485b67e32c947bdf9cd0fcb4b485be718/filelock-4.2.10-py3-none-any.whl
hash:
- sha256: a379c50a729e5b68c8519a4d26489f755d9769960a236ecd838190e2a78db2e9
+ sha256: 9479aaeb4db20b5a53e3da28493edd24720b66b8a0b84030aa56ccdf243d6857
category: main
optional: false
- name: future
@@ -6112,22 +6118,23 @@ package:
category: main
optional: false
- name: liana
- version: 1.7.1
+ version: 2.1.1
manager: pip
platform: linux-64
dependencies:
- anndata: '>=0.7.4'
- docrep: '>=0.3.1'
- mudata: '*'
+ anndata: '>=0.13'
+ fast-array-utils: '*'
+ mudata: '>=0.4'
numba: '*'
pandas: '*'
plotnine: '>=0.10.1'
- scanpy: '*'
+ scanpy: '>=1.12'
+ scverse-misc: '>=0.1.4'
session-info2: '*'
- tqdm: '>=4.0'
- url: https://files.pythonhosted.org/packages/11/35/24bcb5134ad91fb9787c85c0c1ef5f81b1b87adf9dfb29e8edb5e0050ca2/liana-1.7.1-py3-none-any.whl
+ tqdm: '>=4'
+ url: https://files.pythonhosted.org/packages/97/8c/62c88ea51e543cd386cad54de71070322d4a20ed6c899a006cd50799ebf1/liana-2.1.1-py3-none-any.whl
hash:
- sha256: f4eead6df3f7fd07648fda77a50b50673270e1fca19030a7af802d57b84eb2e6
+ sha256: 195253f4fd3f9317e6e41aee114bd29d7a3081cdce5a847a947db1c763472864
category: main
optional: false
- name: lightning
@@ -6256,14 +6263,14 @@ package:
category: main
optional: false
- name: numpy-groupies
- version: 0.16.3
+ version: 0.16.5
manager: pip
platform: linux-64
dependencies:
numpy: '*'
- url: https://files.pythonhosted.org/packages/d6/e2/08809106723ddb004e7af0399b36848667a9f58b40bb85008f5dfc5752d1/numpy_groupies-0.16.3-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/30/75/c7638e6bc615c5bbd044c2e055bb112545767873753546a84d5fb60136e8/numpy_groupies-0.16.5-py3-none-any.whl
hash:
- sha256: 6dd37aff8dc1a172d00589ea93f05d398df9f53f7370266208875ca21542cafb
+ sha256: a9e6cc17d2b93f40ce0d07348091623ee1f5b8c531177721301cb24b602afbfd
category: main
optional: false
- name: opt-einsum
@@ -6344,6 +6351,20 @@ package:
sha256: 36f4526235b3ddd9ea379dd677093ea2ee1b58ab98adec1bde891930283448cd
category: main
optional: false
+- name: plottable
+ version: 0.1.5
+ manager: pip
+ platform: linux-64
+ dependencies:
+ matplotlib: '*'
+ numpy: '*'
+ pandas: '*'
+ pillow: '*'
+ url: https://files.pythonhosted.org/packages/ca/cf/23ea4b6f9460312604ed67b74e7db3e702fcaefc04743224ef1cdda663f4/plottable-0.1.5-py3-none-any.whl
+ hash:
+ sha256: 5c5d1128d90065fc940e5c8445b2603d253909745bb0cee0c9828651b812441c
+ category: main
+ optional: false
- name: progressbar2
version: 4.7.0
manager: pip
@@ -6474,6 +6495,43 @@ package:
sha256: 71f95c7b22f25c4f91cc659e1734bcdc6c69c8b66543f2ed2d78dbe29ce2f167
category: main
optional: false
+- name: scib-metrics
+ version: 0.5.10
+ manager: pip
+ platform: linux-64
+ dependencies:
+ anndata: '*'
+ chex: '*'
+ igraph: '>0.9.0'
+ jax: '*'
+ jaxlib: '*'
+ matplotlib: '*'
+ numpy: '*'
+ pandas: '*'
+ plottable: '*'
+ pynndescent: '*'
+ rich: '*'
+ scanpy: '>=1.9'
+ scikit-learn: '*'
+ scipy: '*'
+ tqdm: '*'
+ umap-learn: '>=0.5.0'
+ url: https://files.pythonhosted.org/packages/fd/66/268cc406d352be0286364bcbb38b9e3c1adbcdea41b37e9e181eb2e4bf32/scib_metrics-0.5.10-py3-none-any.whl
+ hash:
+ sha256: 548639c6f43a8811a10b3cd60aa18ed4d382a13ef68945911e95a4dc063bdf9c
+ category: main
+ optional: false
+- name: scikit-misc
+ version: 0.5.3
+ manager: pip
+ platform: linux-64
+ dependencies:
+ numpy: '>=1.26.4'
+ url: https://files.pythonhosted.org/packages/ec/52/3ba0bb75fe3d9a6d2edbe82463dfbc05f5f665ffa28fb4f123083a673071/scikit_misc-0.5.3-cp312-cp312-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl
+ hash:
+ sha256: 1861b2fa1619dd6645154cdb7d9696d71f0a0e7ef8f5201f1e7bacf21507c858
+ category: main
+ optional: false
- name: scvelo
version: 0.3.4
manager: pip
@@ -6495,32 +6553,33 @@ package:
category: main
optional: false
- name: scvi-tools
- version: 1.4.3
+ version: 1.5.2
manager: pip
platform: linux-64
dependencies:
anndata: '*'
docrep: '>=0.3.2'
- lightning: '>=2.0'
+ joblib: '*'
+ lightning: '>=2'
ml-collections: '*'
mudata: '*'
- numba: '>=0.60.0'
+ numba: '>=0.60'
numpy: '*'
pandas: '*'
pyro-ppl: '*'
rich: '*'
- scanpy: '>=1.10'
+ scanpy: '>=1.11'
scikit-learn: '*'
scipy: '*'
- sparse: '>=0.14.0'
+ sparse: '>=0.14'
tensorboard: '*'
torch: '*'
torchmetrics: '*'
tqdm: '*'
xarray: '*'
- url: https://files.pythonhosted.org/packages/ed/87/ba97e56c8aa0e07ac68a458df5dae9c696cd1a5ffe2b80c9259bb8f7a5d6/scvi_tools-1.4.3-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/89/2c/b672e659ee21e6691c12d6206e1f5c1a6e49ee96c549e77adaf312178b34/scvi_tools-1.5.2-py3-none-any.whl
hash:
- sha256: 635a21db462d27731bc3b1e569203cdfe491ca656546da2fbf139b5574a54960
+ sha256: ecc33564f325b3e6053976725f7eb242060edbbfbc2d9cf13b100405eca790c9
category: main
optional: false
- name: sparse
@@ -6603,15 +6662,25 @@ package:
category: main
optional: false
- name: tensorly
- version: 0.9.0
+ version: 0.10.0
manager: pip
platform: linux-64
dependencies:
numpy: '*'
scipy: '*'
- url: https://files.pythonhosted.org/packages/c3/ad/83380fb3c85ddf68c509059cf328374c1878f87382070475abc72867c7ba/tensorly-0.9.0-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/17/02/0269a1573b9f2ea6365e12b87debaf8b2cdad8e3fc4f84d9f4dc6be68adb/tensorly-0.10.0-py3-none-any.whl
+ hash:
+ sha256: 90becb3d28023638ec60d19fbf8d87735b4ce1537fe3eb881897d69e107b882d
+ category: main
+ optional: false
+- name: toolz
+ version: 1.2.0
+ manager: pip
+ platform: linux-64
+ dependencies: {}
+ url: https://files.pythonhosted.org/packages/db/17/4c8beb6c8c4176c6bf143bfd7e1e4dd6719b00ced90738c7ac471b71c1df/toolz-1.2.0-py3-none-any.whl
hash:
- sha256: a393022ff1616a36c021496f46a955d94234d6bc8c42793f477a6be8849f4873
+ sha256: 890f820b1cb8152785aaf9386d8707770110809035800985ca65cb24ce1120ef
category: main
optional: false
- name: torch
diff --git a/envs/cellquorum-core.yml b/envs/cellquorum-core.yml
index 2cd33841..221f6819 100644
--- a/envs/cellquorum-core.yml
+++ b/envs/cellquorum-core.yml
@@ -12,35 +12,33 @@ dependencies:
- cython
# Core numerical and statistical stack.
- - numpy>=1.24
+ - numpy=2.3.5
- scipy=1.16.3
- - pandas>=2.1
+ - pandas=3.1.1
- scikit-learn>=1.4
- statsmodels>=0.14
- numba>=0.59
# Single-cell data structures and analysis.
- - anndata>=0.12,<0.13
+ - anndata=0.13.4
# mudata (a scvi-tools and liana dependency) imports Settings from scverse-misc, which needs scverse-misc's
# `settings` extra: pydantic-settings and python-dotenv. A conda install of scverse-misc does not bring
# extras, so without these two `import mudata`, `import scvi` and `import liana` fail.
- pydantic-settings
- python-dotenv
- - scanpy>=1.10
+ - pooch>=1.8,<2
+ - scanpy=1.12.5
- h5py>=3.10
- zarr>=2.17
- natsort>=8.4
# Neighbor graphs, dimensionality reduction, and clustering.
- - umap-learn>=0.5
+ - umap-learn=0.5.12
- pynndescent>=0.5
- python-igraph>=0.11
- leidenalg>=0.10
- louvain>=0.8
- # HVG and preprocessing support.
- - scikit-misc>=0.3
-
# CellRank sparse Schur solver (SLEPc/PETSc), as CONDA packages — building
# these from pip is fragile (PETSc configures + compiles from source). Without
# them, CellRank's GPCCA Schur decomposition silently falls back to the dense
@@ -60,7 +58,7 @@ dependencies:
- scrublet>=0.2
# Biological activity and molecular inference.
- - decoupler-py>=1.8
+ - decoupler-py=2.2.0
# Network/statistical utilities.
- networkx>=3.2
@@ -99,7 +97,7 @@ dependencies:
# this env; the cellquorum-gpu env installs the CUDA build.
- pip:
- torch @ https://download-r2.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-manylinux_2_28_x86_64.whl#sha256=cb9a8ba8137ab24e36bf1742cb79a1294bd374db570f09fc15a5e1318160db4e
- - anndata==0.12.19
+ - scikit-misc==0.5.3
- pygpcca @ git+https://github.com/msmdev/pygpcca.git@773c463cc5043a5d40e6f31bbbc20f4e49587008
# NOTE: the editable install of the repo (`pip install -e .`) is NOT done
# here — at env-create time the source is not yet copied into the image,
@@ -107,15 +105,19 @@ dependencies:
# the image), where no package exists. The Dockerfile installs the package
# editable in a dedicated RUN step after `COPY . /opt/cellquorum`.
# Cell-cell communication (in-process).
- - liana==1.7.1
- - cell2cell==0.8.4
- - tensorly==0.9.0
+ - liana==2.1.1
+ - cell2cell==0.9.0
+ - tensorly==0.10.0
# Trajectory / RNA velocity / potency (in-process).
- scvelo==0.3.4
- - cellrank==2.3.2
+ - cellrank==2.3.3
- palantir==1.4.4
# Deep generative models (scVI/scArches); pulls torch.
- - scvi-tools==1.4.3
+ - scvi-tools==1.5.2
+ # Integration quality on the two-study human pancreas reference.
+ - scib-metrics==0.5.10
+ - chex==0.1.90
+ - plottable==0.1.5
# Imported lazily by stages that otherwise skip or fail without them: PHATE embeddings, CellTypist
# annotation, and the chord and Sankey communication figures. They worked on a workstation because
# someone installed them by hand, and were missing from every fresh install and the container.
diff --git a/envs/cellquorum-gpu.conda-lock.yml b/envs/cellquorum-gpu.conda-lock.yml
index 91897a11..43deb083 100644
--- a/envs/cellquorum-gpu.conda-lock.yml
+++ b/envs/cellquorum-gpu.conda-lock.yml
@@ -13,14 +13,10 @@
version: 1
metadata:
content_hash:
- linux-64: 6c0b247b1131d1b5bcd30a34e995efd05633e4da70131582a328a099b24ef7a7
+ linux-64: ef6198adf415a87d61855e0fbf32d269ea61d1f19911128d5c1f03cb3ff5600a
channels:
- - url: rapidsai
- used_env_vars: []
- url: conda-forge
used_env_vars: []
- - url: nvidia
- used_env_vars: []
- url: bioconda
used_env_vars: []
platforms:
@@ -68,55 +64,6 @@ package:
sha256: 5a5332632469b3a8babc1bdcbaccfd7c2a6ec248e46723d14a35e4cd2d684330
category: main
optional: false
-- name: aiohappyeyeballs
- version: 2.7.1
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.10'
- url: https://conda.anaconda.org/conda-forge/noarch/aiohappyeyeballs-2.7.1-pyhd8ed1ab_0.conda
- hash:
- md5: e51e09bf3b91c62b7461a9f94e92c2c9
- sha256: 5ef589e5fd3736439781a9d48e68ce1e723b7081a471c02169908198eba4f448
- category: main
- optional: false
-- name: aiohttp
- version: 3.14.5
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- aiohappyeyeballs: '>=2.5.0'
- aiosignal: '>=1.4.0'
- attrs: '>=17.3.0'
- frozenlist: '>=1.1.1'
- libgcc: '>=15'
- multidict: '>=4.5,<8.0'
- propcache: '>=0.2.0'
- python: ''
- python_abi: 3.12.*
- typing_extensions: '>=4.4'
- yarl: '>=1.25.1,<2.0'
- url: https://conda.anaconda.org/conda-forge/linux-64/aiohttp-3.14.5-py312h8b7ec3f_0.conda
- hash:
- md5: e5e9629c0a1c75dc79ec1ac35b0560da
- sha256: 5d1da5a10fa2e296cbfd345cca3a0f82d985e480e14ec40fe206ddb777dc3a2b
- category: main
- optional: false
-- name: aiosignal
- version: 1.4.0
- manager: conda
- platform: linux-64
- dependencies:
- frozenlist: '>=1.1.0'
- python: '>=3.9'
- typing_extensions: '>=4.2'
- url: https://conda.anaconda.org/conda-forge/noarch/aiosignal-1.4.0-pyhd8ed1ab_0.conda
- hash:
- md5: 421a865222cd0c9d83ff08bc78bf3a61
- sha256: 8dc149a6828d19bf104ea96382a9d04dae185d4a03cc6beb1bc7b84c428e3ca2
- category: main
- optional: false
- name: alsa-lib
version: 1.2.16.1
manager: conda
@@ -131,37 +78,40 @@ package:
category: main
optional: false
- name: anndata
- version: 0.12.19
+ version: 0.13.4
manager: conda
platform: linux-64
dependencies:
array-api-compat: '>=1.7.1'
- h5py: '>=3.8'
+ h5py: '>=3.11'
legacy-api-wrap: ''
natsort: ''
- numpy: '>=1.26'
+ numpy: '>=2.1'
packaging: '>=24.2'
- pandas: '>=2.1.0,!=2.1.2,<3'
- python: '>=3.11'
- scipy: '>=1.12,!=1.17.0'
- scverse-misc: '>=0.0.3'
- zarr: '>=2.18.7,!=3.0.*'
- url: https://conda.anaconda.org/conda-forge/noarch/anndata-0.12.19-pyhd8ed1ab_0.conda
+ pandas: '>=2.3'
+ pydantic-settings: ''
+ python: '>=3.12'
+ python-dotenv: ''
+ scipy: '>=1.14,!=1.17.0'
+ scverse-misc: '>=0.1.0'
+ typing_extensions: '>=4.16'
+ zarr: '>=3.1'
+ url: https://conda.anaconda.org/conda-forge/noarch/anndata-0.13.4-pyhd8ed1ab_0.conda
hash:
- md5: cb6497a1f1f2e3f83256e8eadb995bd2
- sha256: f9253f582480d2f49e8b1ca49922ccdaa1815674bd2b83742c11975f1116a001
+ md5: 8c7b633e969cec15952d4318d7cbb66f
+ sha256: c325ebd706c72e7db772f5311f67373cbc2b1ad35b0e3fe0c51f803b8e8a41d6
category: main
optional: false
- name: annotated-doc
- version: 0.0.5
+ version: 0.0.6
manager: conda
platform: linux-64
dependencies:
- python: '>=3.10'
- url: https://conda.anaconda.org/conda-forge/noarch/annotated-doc-0.0.5-pyhcf101f3_0.conda
+ python: '>=3.11'
+ url: https://conda.anaconda.org/conda-forge/noarch/annotated-doc-0.0.6-pyh5ded981_0.conda
hash:
- md5: d68ec17cb915cab4642357417ec0a104
- sha256: a0a320c709fded9b5178108dedebf81a1f5a5d9c7720e3af50c1ab058dadd296
+ md5: 9fca85fe0c3f76ab2199f576d7d115c0
+ sha256: ade8040829ed573c628afb85868a74bd03df7f3c663be318f6e5bbad2e905f78
category: main
optional: false
- name: annotated-types
@@ -249,18 +199,6 @@ package:
sha256: fbbd8ce60cbd5c16f3fe559eb644551f94285caff30985ea961ff851c1cf25ac
category: main
optional: false
-- name: attrs
- version: 26.1.0
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.10'
- url: https://conda.anaconda.org/conda-forge/noarch/attrs-26.1.0-pyhcf101f3_0.conda
- hash:
- md5: c6b0543676ecb1fb2d7643941fe375f2
- sha256: 1b6124230bb4e571b1b9401537ecff575b7b109cc3a21ee019f65e083b8399ab
- category: main
- optional: false
- name: aws-c-auth
version: 1.0.0
manager: conda
@@ -599,27 +537,6 @@ package:
sha256: 6d9aeebe7a06acc11f52c6a98046a9ffffac8c2179cdfb918a161467139156f3
category: main
optional: false
-- name: bokeh
- version: 3.10.0
- manager: conda
- platform: linux-64
- dependencies:
- contourpy: '>=1.2'
- jinja2: '>=2.9'
- narwhals: '>=1.13'
- numpy: '>=1.16'
- packaging: '>=16.8'
- pillow: '>=7.1.0'
- python: '>=3.12'
- pyyaml: '>=3.10'
- tornado: '>=6.2'
- xyzservices: '>=2021.09.1'
- url: https://conda.anaconda.org/conda-forge/noarch/bokeh-3.10.0-pyhd8ed1ab_0.conda
- hash:
- md5: 67d0ed593951a4933024453b76ef1355
- sha256: 34bb5cc4f6ccff1e410dea2d39880190f88a8be7ad5cfe9ff0da2cf3d2ad7ea1
- category: main
- optional: false
- name: brotli
version: 1.2.1
manager: conda
@@ -729,18 +646,6 @@ package:
sha256: 4bd96dab5a434e4ee59fb6910cae704d4c5b453907d4886272a9943d37e42b6f
category: main
optional: false
-- name: cachetools
- version: 7.3.0
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.11'
- url: https://conda.anaconda.org/conda-forge/noarch/cachetools-7.3.0-pyhd8ed1ab_0.conda
- hash:
- md5: 3bcfa89219c46a5b7c047eca5ece20bf
- sha256: bca3b83480983877e94e771141a77974721fa298f343bd79de04af96bdc98b82
- category: main
- optional: false
- name: cairo
version: 1.18.6
manager: conda
@@ -800,18 +705,6 @@ package:
sha256: e4f04b45d68716fcc47c808aeacbf165f6f295fb91bd20b91aa38942b8d2cdc6
category: main
optional: false
-- name: click
- version: 8.5.0
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.11'
- url: https://conda.anaconda.org/conda-forge/noarch/click-8.5.0-pyh5ded981_0.conda
- hash:
- md5: dea5b13a211bbf99876deb980b408a66
- sha256: 9afb0c2c089330321a219c7ee5a25313bd574d4024be606cbb33e7a5735df662
- category: main
- optional: false
- name: cloudpickle
version: 3.1.2
manager: conda
@@ -884,505 +777,14 @@ package:
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libgcc: '>=15'
- libstdcxx: '>=15'
- python: ''
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/cramjam-2.14.0-py312h415a56e_1.conda
- hash:
- md5: c26c18257b89dc047863c07356bb8618
- sha256: c2b5a13e1858902db298129ed95ebfc119ae1149cb8bac7655cf1166a240319a
- category: main
- optional: false
-- name: cuda-bindings
- version: 12.9.9
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-nvcc-impl: '>=12,<13.0a0'
- cuda-nvrtc: '>=12,<13.0a0'
- cuda-pathfinder: '>=1.5.5,<2'
- cuda-version: '>=12,<13.0a0'
- libcufile: '>=1,<2.0a0'
- libgcc: '>=15'
- libnvjitlink: '>=12.3,<13'
- libstdcxx: '>=15'
- python: ''
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-bindings-12.9.9-py312h3956a73_1.conda
- hash:
- md5: 120c43d21aa34beefb88f977290471c6
- sha256: 8f0f5aec3e60b8877e57b473983ebc18e7ee2c0ec759a77af8b180abd02b48b9
- category: main
- optional: false
-- name: cuda-cccl_linux-64
- version: 12.9.27
- manager: conda
- platform: linux-64
- dependencies:
- cuda-version: '>=12.9,<12.10.0a0'
- url: https://conda.anaconda.org/conda-forge/noarch/cuda-cccl_linux-64-12.9.27-ha770c72_0.conda
- hash:
- md5: 87ff6381e33b76e5b9b179a2cdd005ec
- sha256: 2ee3b9564ca326226e5cda41d11b251482df8e7c757e333d28ec75213c75d126
- category: main
- optional: false
-- name: cuda-core
- version: 0.7.0
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-bindings: '>=12,<13.0a0'
- cuda-pathfinder: '>=1.4.2,<2'
- cuda-version: '>=12,<13.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- numpy: ''
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-core-0.7.0-cuda12_py312h32364ff_0.conda
- hash:
- md5: f5909d4676ce048eba25c6d41ddf9aae
- sha256: 0f30b78b1d0a02194a537166e2cec6e92a09e8c60464b951e58686eee8b802c5
- category: main
- optional: false
-- name: cuda-crt-dev_linux-64
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- cuda-version: '>=12.9,<12.10.0a0'
- url: https://conda.anaconda.org/conda-forge/noarch/cuda-crt-dev_linux-64-12.9.86-ha770c72_2.conda
- hash:
- md5: 79d280de61e18010df5997daea4743df
- sha256: e6257534c4b4b6b8a1192f84191c34906ab9968c92680fa09f639e7846a87304
- category: main
- optional: false
-- name: cuda-crt-tools
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- cuda-version: '>=12.9,<12.10.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-crt-tools-12.9.86-ha770c72_2.conda
- hash:
- md5: 503a94e20d2690d534d676a764a1852c
- sha256: 2da9964591af14ba11b2379bed01d56e7185260ee0998d1a939add7fb752db45
- category: main
- optional: false
-- name: cuda-cudart
- version: 12.9.79
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-cudart_linux-64: 12.9.79
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=13'
- libstdcxx: '>=13'
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-cudart-12.9.79-h5888daf_0.conda
- hash:
- md5: cb15315d19b58bd9cd424084e58ad081
- sha256: 57d1294ecfaf9dc8cdb5fc4be3e63ebc7614538bddb5de53cfd9b1b7de43aed5
- category: main
- optional: false
-- name: cuda-cudart-dev
- version: 12.9.79
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-cudart: 12.9.79
- cuda-cudart-dev_linux-64: 12.9.79
- cuda-cudart-static: 12.9.79
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=13'
- libstdcxx: '>=13'
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-cudart-dev-12.9.79-h5888daf_0.conda
- hash:
- md5: ba38a7c3b4c14625de45784b773f0c71
- sha256: 04d8235cb3cb3510c0492c3515a9d1a6053b50ef39be42b60cafb05044b5f4c6
- category: main
- optional: false
-- name: cuda-cudart-dev_linux-64
- version: 12.9.79
- manager: conda
- platform: linux-64
- dependencies:
- cuda-cccl_linux-64: ''
- cuda-cudart-static_linux-64: ''
- cuda-cudart_linux-64: ''
- cuda-version: '>=12.9,<12.10.0a0'
- url: https://conda.anaconda.org/conda-forge/noarch/cuda-cudart-dev_linux-64-12.9.79-h3f2d84a_0.conda
- hash:
- md5: 86e40eb67d83f1a58bdafdd44e5a77c6
- sha256: ffe86ed0144315b276f18020d836c8ef05bf971054cf7c3eb167af92494080d5
- category: main
- optional: false
-- name: cuda-cudart-static
- version: 12.9.79
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-cudart-static_linux-64: 12.9.79
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=13'
- libstdcxx: '>=13'
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-cudart-static-12.9.79-h5888daf_0.conda
- hash:
- md5: d3c4ac48f4967f09dd910d9c15d40c81
- sha256: 6261e1d9af80e1ec308e3e5e2ff825d189ef922d24093beaf6efca12e67ce060
- category: main
- optional: false
-- name: cuda-cudart-static_linux-64
- version: 12.9.79
- manager: conda
- platform: linux-64
- dependencies:
- cuda-version: '>=12.9,<12.10.0a0'
- url: https://conda.anaconda.org/conda-forge/noarch/cuda-cudart-static_linux-64-12.9.79-h3f2d84a_0.conda
- hash:
- md5: b87bf315d81218dd63eb46cc1eaef775
- sha256: d435f8a19b59b52ce460ee3a6bfd877288a0d1d645119a6ba60f1c3627dc5032
- category: main
- optional: false
-- name: cuda-cudart_linux-64
- version: 12.9.79
- manager: conda
- platform: linux-64
- dependencies:
- cuda-version: '>=12.9,<12.10.0a0'
- url: https://conda.anaconda.org/conda-forge/noarch/cuda-cudart_linux-64-12.9.79-h3f2d84a_0.conda
- hash:
- md5: 64508631775fbbf9eca83c84b1df0cae
- sha256: 6cde0ace2b995b49d0db2eefb7bc30bf00ffc06bb98ef7113632dec8f8907475
- category: main
- optional: false
-- name: cuda-nvcc-dev_linux-64
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- cuda-crt-dev_linux-64: 12.9.86
- cuda-nvvm-dev_linux-64: 12.9.86
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=6'
- libnvptxcompiler-dev_linux-64: 12.9.86
- url: https://conda.anaconda.org/conda-forge/noarch/cuda-nvcc-dev_linux-64-12.9.86-he91c749_2.conda
- hash:
- md5: 19d4e090217f0ea89d30bedb7461c048
- sha256: a1672a34439a72869de9e011e935d41b62fc8dfb1a2700e85ed8a7a129b79981
- category: main
- optional: false
-- name: cuda-nvcc-impl
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- cuda-cudart: '>=12.9.79,<13.0a0'
- cuda-cudart-dev: ''
- cuda-nvcc-dev_linux-64: 12.9.86
- cuda-nvcc-tools: 12.9.86
- cuda-nvvm-impl: 12.9.86
- cuda-version: '>=12.9,<12.10.0a0'
- libnvptxcompiler-dev: 12.9.86
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-nvcc-impl-12.9.86-h85509e4_2.conda
- hash:
- md5: 67458d2685e7503933efa550f3ee40f3
- sha256: 961cf20d411b7685cd744e6c6ed35efea547d095c62151d6f3053d9931bb994d
- category: main
- optional: false
-- name: cuda-nvcc-tools
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-crt-tools: 12.9.86
- cuda-nvvm-tools: 12.9.86
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=12'
- libstdcxx: '>=12'
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-nvcc-tools-12.9.86-he02047a_2.conda
- hash:
- md5: dc256c9864c2e8e9c817fbca1c84a4bc
- sha256: 0e849be7b5e4832ca218ec2c48a9ba3a15a984f629e2e54f38a53f4f57220341
- category: main
- optional: false
-- name: cuda-nvrtc
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-nvrtc-12.9.86-hecca717_1.conda
- hash:
- md5: 53f0062e2243b26e43ddac0b5267c6a3
- sha256: 68f81268c25befa9b70dc49af469ab0eb131960e3700b9a4edb46a32da343a28
- category: main
- optional: false
-- name: cuda-nvvm-dev_linux-64
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- cuda-version: '>=12.9,<12.10.0a0'
- url: https://conda.anaconda.org/conda-forge/noarch/cuda-nvvm-dev_linux-64-12.9.86-ha770c72_2.conda
- hash:
- md5: 7b386291414c7eea113d25ac28a33772
- sha256: 522722dcaffd133e0c7500c69dc70e21ac34d6762dcbaabfe847439f944028f0
- category: main
- optional: false
-- name: cuda-nvvm-impl
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=12'
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-nvvm-impl-12.9.86-h4bc722e_2.conda
- hash:
- md5: 82125dd3c0c4aa009faa00e2829b93d8
- sha256: f4d34556174e4faa9d374ba2244707082870e1bbc1bb441ad3d9d2cea37da6af
- category: main
- optional: false
-- name: cuda-nvvm-tools
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=12'
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-nvvm-tools-12.9.86-h4bc722e_2.conda
- hash:
- md5: f9af26e4079adcd72688a8e8dbecb229
- sha256: 45f5e881ed0d973132a5475a0b5c066db6e748ef3a831a14dba8374b252e0067
- category: main
- optional: false
-- name: cuda-pathfinder
- version: 1.8.6
- manager: conda
- platform: linux-64
- dependencies:
- cuda-version: '>=12.0,<14'
- python: '>=3.11'
- url: https://conda.anaconda.org/conda-forge/noarch/cuda-pathfinder-1.8.6-pyh5ded981_0.conda
- hash:
- md5: 23edebeb1f3f71dc1b5341e822c99894
- sha256: b6f19dbea229dd326b76c7d89ce999f81426c4a5103ae97465986c4d1db36eec
- category: main
- optional: false
-- name: cuda-profiler-api
- version: 12.9.79
- manager: conda
- platform: linux-64
- dependencies:
- cuda-cudart-dev: ''
- cuda-version: '>=12.9,<12.10.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/cuda-profiler-api-12.9.79-h7938cbb_1.conda
- hash:
- md5: 90d09865fb37d11d510444e34ebe6a09
- sha256: 4f679dfbf2bf2d17abb507f31b0176c0e3572337b5005b9e36179948a53988ac
- category: main
- optional: false
-- name: cuda-python
- version: 12.9.9
- manager: conda
- platform: linux-64
- dependencies:
- cuda-bindings: '>=12.9.9,<12.10.0a0'
- cuda-version: '>=12.0,<13.0a0'
- python: '>=3.10'
- url: https://conda.anaconda.org/conda-forge/noarch/cuda-python-12.9.9-py_min_310_1.conda
- hash:
- md5: 37b78877c0101c248fe13ff9d4c84585
- sha256: 40ea0297de48f299f6361fa9a840c4713a9a769b58b3feebd7b52f2a03547d28
- category: main
- optional: false
-- name: cuda-version
- version: '12.9'
- manager: conda
- platform: linux-64
- dependencies: {}
- url: https://conda.anaconda.org/conda-forge/noarch/cuda-version-12.9-h4f385c5_3.conda
- hash:
- md5: b6d5d7f1c171cbd228ea06b556cfa859
- sha256: 5f5f428031933f117ff9f7fcc650e6ea1b3fef5936cf84aa24af79167513b656
- category: main
- optional: false
-- name: cudf
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- _python_abi3_support: 1.*
- cachetools: ''
- cpython: '>=3.11'
- cuda-cudart: ''
- cuda-python: '>=12.9.2,<13.0'
- cuda-version: '>=12,<13.0a0'
- cupy: '>=13.6.0,!=14.0.0'
- fsspec: '>=0.6.0'
- libcudf: 26.4.0.*,>=26.4.0,<26.5.0a0
- libcufile: ''
- libgcc: '>=14'
- libstdcxx: '>=14'
- numba: '>=0.60.0,<0.65.0'
- numba-cuda: '>=0.22.2,<0.29.0'
- numpy: '>=1.26,<3.0'
- nvtx: '>=0.2.1'
- packaging: ''
- pandas: '>=2.0,<2.4.0'
- pyarrow: '>=19.0.0'
- pylibcudf: 26.4.0.*
- python: ''
- rich: ''
- rmm: '>=26.4.0,<26.5.0a0'
- url: https://conda.anaconda.org/rapidsai/linux-64/cudf-26.04.00-cuda12_cp311_abi3_260408_f9c3cf19.conda
- hash:
- md5: 57f91cdd900f08e2ada54976109c9a0e
- sha256: dd625bdb0931a88d7f4edf2252cf5f009b142ca75a0a1a7f9681d501fb1d6e2a
- category: main
- optional: false
-- name: cugraph
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- _python_abi3_support: 1.*
- aiohttp: ''
- cpython: '>=3.11'
- cuda-python: '>=12.9.2,<13.0'
- cuda-version: '>=12,<13.0a0'
- cudf: 26.4.*
- cupy: '>=13.6.0'
- dask-cuda: 26.4.*
- dask-cudf: 26.4.*
- fsspec: '>=0.6.0'
- libcugraph: 26.4.0.*
- libgcc: '>=14'
- libstdcxx: '>=14'
- numba: '>=0.60.0,<0.65.0'
- numba-cuda: '>=0.22.2'
- pylibcudf: 26.4.*
- pylibcugraph: 26.4.0.*
- pylibraft: 26.4.*
- python: ''
- raft-dask: 26.4.*
- rapids-dask-dependency: 26.4.*
- requests: ''
- ucxx: 0.49.*
- url: https://conda.anaconda.org/rapidsai/linux-64/cugraph-26.04.00-cuda12_cp311_abi3_260408_03ba9c2d.conda
- hash:
- md5: 3a30e2ac0e277b50feb0e46ddffe6b3b
- sha256: a598774d6a6fd2c08744d5d6333a23d12e9917f446d25f992818b82199925f92
- category: main
- optional: false
-- name: cuml
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- _python_abi3_support: 1.*
- cpython: '>=3.11'
- cuda-cudart: ''
- cuda-python: '>=12.9.2,<13.0'
- cuda-version: '>=12,<13.0a0'
- cudf: 26.4.*
- cupy: '>=13.6.0'
- joblib: '>=0.11'
- libcuml: 26.4.0.*
- libgcc: '>=14'
- libstdcxx: '>=14'
- numba: '>=0.60.0,<0.65.0'
- numba-cuda: '>=0.22.2,<0.29.0'
- numpy: '>=1.23,<3.0'
- packaging: ''
- pylibraft: 26.4.*
- python: ''
- rapids-logger: 0.2.*
- rich: ''
- scikit-learn: '>=1.4'
- scipy: '>=1.14.0'
- treelite: '>=4.7.0,<5.0.0'
- url: https://conda.anaconda.org/rapidsai/linux-64/cuml-26.04.00-cuda12_cp311_abi3_260409_f223d89b.conda
- hash:
- md5: 198724faac79606a86a81aeec5b95b57
- sha256: f9d2f3fd24fedb32cb8aeddb002f403907ab681ab45627b945b9aa534278e947
- category: main
- optional: false
-- name: cupy
- version: 14.1.1
- manager: conda
- platform: linux-64
- dependencies:
- cuda-cudart-dev_linux-64: ''
- cuda-nvrtc: ''
- cuda-version: '>=12,<13.0a0'
- cupy-core: 14.1.1
- libcublas: ''
- libcufft: ''
- libcurand: ''
- libcusolver: ''
- libcusparse: ''
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/cupy-14.1.1-py312h53aca36_0.conda
- hash:
- md5: 7da0de2f5134adb58428dd058e8687b2
- sha256: 31bca442aa1731226acbf43293803eedfc3a4c4a18930aeec46f1f82faaeff2b
- category: main
- optional: false
-- name: cupy-core
- version: 14.1.1
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-pathfinder: '>=1.3.4,<2.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- numpy: '>=1.23,<3,>=2.0'
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/cupy-core-14.1.1-py312h16a6543_0.conda
- hash:
- md5: 794c6c5a4654f45265339382cdc6d28d
- sha256: 0933f9fb3f542799ec06efdb1c6c925800969cc4fe76574df6cdf18d360d4ed2
- category: main
- optional: false
-- name: cuvs
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- _python_abi3_support: 1.*
- cpython: '>=3.11'
- cuda-cudart: ''
- cuda-python: '>=12.9.2,<13.0'
- cuda-version: '>=12,<13.0a0'
- libcuvs: 26.4.0.*
- libgcc: '>=14'
- libstdcxx: '>=14'
- numpy: '>=1.23,<3.0'
- pylibraft: 26.4.*
+ libgcc: '>=15'
+ libstdcxx: '>=15'
python: ''
- url: https://conda.anaconda.org/rapidsai/linux-64/cuvs-26.04.00-cuda12_cp311_abi3_260408_.conda
+ python_abi: 3.12.*
+ url: https://conda.anaconda.org/conda-forge/linux-64/cramjam-2.14.0-py312h415a56e_1.conda
hash:
- md5: a1a9e67d1b2f9134c41cb4a4063c7497
- sha256: 2b4475525e4f49e338450346ee263b24b146c5be59124ac51aa674acfdf922de
+ md5: c26c18257b89dc047863c07356bb8618
+ sha256: c2b5a13e1858902db298129ed95ebfc119ae1149cb8bac7655cf1166a240319a
category: main
optional: false
- name: cycler
@@ -1431,99 +833,6 @@ package:
sha256: 08eae1f2a25f8bfe6e50feedc92019a729dd3a2d2419f0acb14e51768af9f1fa
category: main
optional: false
-- name: cytoolz
- version: 1.2.0
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=15'
- python: ''
- python_abi: 3.12.*
- toolz: '>=0.10.0'
- url: https://conda.anaconda.org/conda-forge/linux-64/cytoolz-1.2.0-py312h1b36aeb_0.conda
- hash:
- md5: 20f3a25d978a201f5dc592380781a7e6
- sha256: 2da7f1f86606e8501ceb1d5a1f1637e5a324b04fee395878b02c770c56a0fafe
- category: main
- optional: false
-- name: dask
- version: 2026.1.1
- manager: conda
- platform: linux-64
- dependencies:
- bokeh: '>=3.1.0'
- cytoolz: '>=0.11.0'
- dask-core: '>=2026.1.1,<2026.1.2.0a0'
- distributed: '>=2026.1.1,<2026.1.2.0a0'
- jinja2: '>=2.10.3'
- lz4: '>=4.3.2'
- numpy: '>=1.24'
- pandas: '>=2.0'
- pyarrow: '>=14.0.1'
- python: '>=3.10'
- url: https://conda.anaconda.org/conda-forge/noarch/dask-2026.1.1-pyhcf101f3_0.conda
- hash:
- md5: a86541105aa7920d2147d48bf370dc08
- sha256: cc3a106881051c9a4bdaf05fdf7e175f0d63885b4624728101e7996a0e6e8ae3
- category: main
- optional: false
-- name: dask-core
- version: 2026.1.1
- manager: conda
- platform: linux-64
- dependencies:
- click: '>=8.1'
- cloudpickle: '>=3.0.0'
- fsspec: '>=2021.9.0'
- importlib-metadata: '>=4.13.0'
- packaging: '>=20.0'
- partd: '>=1.4.0'
- python: '>=3.10'
- pyyaml: '>=5.3.1'
- toolz: '>=0.12.0'
- url: https://conda.anaconda.org/conda-forge/noarch/dask-core-2026.1.1-pyhcf101f3_1.conda
- hash:
- md5: 91e3b2a0d014ac032c066a2e18051686
- sha256: f279cecdcc132861e49c8f779ff3bfd42b8de811ca97b82566b6c7b23a136b11
- category: main
- optional: false
-- name: dask-cuda
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28'
- __linux: ''
- click: '>=8.1'
- cuda-core: '>=0.3'
- numpy: '>=1.23,<3.0'
- nvidia-ml-py: '>=12'
- pandas: '>=1.3'
- python: ''
- rapids-dask-dependency: 26.4.*
- zict: '>=2.0.0'
- url: https://conda.anaconda.org/rapidsai/noarch/dask-cuda-26.04.00-py314_260408_157ffa32.conda
- hash:
- md5: 4fe93ec1043a6e55608844668f0587ea
- sha256: 0aa3bd13cb7f802d406c833bf1f34a6bcb117deb4d9eabaf7db556e6e5b27fe2
- category: main
- optional: false
-- name: dask-cudf
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- cudf: 26.4.0.*
- nvidia-ml-py: '>=12'
- python: ''
- rapids-dask-dependency: 26.4.*
- url: https://conda.anaconda.org/rapidsai/noarch/dask-cudf-26.04.00-260408_f9c3cf19.conda
- hash:
- md5: 91efe883f9aeb3c9042b332e79fe037d
- sha256: f78455a31793e663ba04cdcc15fd4e536a95c2193ee15bee4db469ef2e8bb44a
- category: main
- optional: false
- name: dbus
version: 1.16.2
manager: conda
@@ -1560,7 +869,7 @@ package:
category: main
optional: false
- name: decoupler-py
- version: 2.3.1
+ version: 2.2.0
manager: conda
platform: linux-64
dependencies:
@@ -1569,72 +878,15 @@ package:
docrep: ''
ipywidgets: ''
marsilea: ''
- python: '>=3.12'
+ python: '>=3.11'
python-igraph: ''
requests: ''
scanpy: ''
xgboost: ''
- url: https://conda.anaconda.org/conda-forge/noarch/decoupler-py-2.3.1-pyhd8ed1ab_0.conda
- hash:
- md5: 5f873c51e540df493e9c21c61dea146c
- sha256: ba9801b7a32aac3e8661e7f68a27bbfc28e7cb57a092a8992b2ee84ee444304d
- category: main
- optional: false
-- name: distributed
- version: 2026.1.1
- manager: conda
- platform: linux-64
- dependencies:
- click: '>=8.0'
- cloudpickle: '>=3.0.0'
- cytoolz: '>=0.12.0'
- dask-core: '>=2026.1.1,<2026.1.2.0a0'
- jinja2: '>=2.10.3'
- locket: '>=1.0.0'
- msgpack-python: '>=1.0.2'
- packaging: '>=20.0'
- psutil: '>=5.8.0'
- python: '>=3.10'
- pyyaml: '>=5.4.1'
- sortedcontainers: '>=2.0.5'
- tblib: '>=1.6.0'
- toolz: '>=0.12.0'
- tornado: '>=6.2.0'
- urllib3: '>=1.26.5'
- zict: '>=3.0.0'
- url: https://conda.anaconda.org/conda-forge/noarch/distributed-2026.1.1-pyhcf101f3_1.conda
- hash:
- md5: c15e359a982395be86a7576a91f9c5f5
- sha256: cc3dc8383c6693e387dba37a7feb360df0c7f3d8c5c11c92f4ca173e9a18476f
- category: main
- optional: false
-- name: distributed-ucxx
- version: 0.49.00
- manager: conda
- platform: linux-64
- dependencies:
- cuda-core: '>=0.3.2'
- python: ''
- pyyaml: '>=6'
- rapids-dask-dependency: 26.4.*
- ucxx: '>=0.49.0,<0.50.0a0'
- url: https://conda.anaconda.org/rapidsai/noarch/distributed-ucxx-0.49.00-260408_8d47a9ff_h546327b.conda
- hash:
- md5: 27adbace1ea2653e42b5c79236fcf392
- sha256: fc916036bb0b6553706b5d9f5288e92ec276d821048bb330350a7f69ad20ffa6
- category: main
- optional: false
-- name: dlpack
- version: '0.8'
- manager: conda
- platform: linux-64
- dependencies:
- libgcc-ng: '>=12'
- libstdcxx-ng: '>=12'
- url: https://conda.anaconda.org/conda-forge/linux-64/dlpack-0.8-h59595ed_3.conda
+ url: https://conda.anaconda.org/conda-forge/noarch/decoupler-py-2.2.0-pyhd8ed1ab_0.conda
hash:
- md5: ee290dba0b7497f2d357c057aaea123f
- sha256: 5884a5e18a779586a2179c0187ef75caa148568dd576c4dbc278deead77cf4b1
+ md5: 47fb2cb7ca1f606ca4e5bc98a96ba597
+ sha256: f20d633d2d42e0a9dbf8d5ec18a70e234d046aba4ea3f68b540ad3756e7dc27e
category: main
optional: false
- name: docrep
@@ -1790,10 +1042,10 @@ package:
manager: conda
platform: linux-64
dependencies: {}
- url: https://conda.anaconda.org/conda-forge/noarch/font-ttf-ubuntu-0.83-h77eed37_3.conda
+ url: https://conda.anaconda.org/conda-forge/noarch/font-ttf-ubuntu-0.83-h77eed37_4.conda
hash:
- md5: 49023d73832ef61042f6a237cb2687e7
- sha256: 2821ec1dc454bd8b9a31d0ed22a7ce22422c0aef163c59f49dfdf915d0f0ca14
+ md5: 1df978a6efeb98fa80b96f2860b9b280
+ sha256: a7384ac883b1acf492fd1aef27431f0f1a647bbb8b5709a1bfab11b0a765f5e5
category: main
optional: false
- name: fontconfig
@@ -1904,22 +1156,6 @@ package:
sha256: 6cabdf62dad4973ef0fe8b0d80c4b5b5d76ac9249da91f26f663cb44acdcc6a1
category: main
optional: false
-- name: frozenlist
- version: 1.8.0
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=15'
- libstdcxx: '>=15'
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/frozenlist-1.8.0-py312hd8b35f9_1.conda
- hash:
- md5: 74f6f936099812a232809e7f607a8af7
- sha256: c2cd7475baceeb6be89e0c7d498df1b393514a69c3b2c5f0ae0abf130001a112
- category: main
- optional: false
- name: fsspec
version: 2026.9.0
manager: conda
@@ -2562,6 +1798,7 @@ package:
azure-core-cpp: '>=1.16.4,<1.16.5.0a0'
azure-identity-cpp: '>=1.13.3,<1.13.4.0a0'
azure-storage-blobs-cpp: '>=12.19.0,<12.19.1.0a0'
+ azure-storage-common-cpp: '>=12.15.0,<12.15.1.0a0'
azure-storage-files-datalake-cpp: '>=12.17.0,<12.17.1.0a0'
bzip2: '>=1.0.8,<2.0a0'
glog: '>=0.7.1,<0.8.0a0'
@@ -2578,13 +1815,13 @@ package:
lz4-c: '>=1.10.0,<1.11.0a0'
orc: '>=2.3.1,<2.3.2.0a0'
simdjson: '>=5.1.1,<5.2.0a0'
- snappy: '>=1.2.2,<1.3.0a0'
+ snappy: '>=1.3.1,<1.4.0a0'
uriparser: '>=0.9.8,<1.0a0'
zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-26.0.1-hb81f139_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-26.0.1-h4f621f8_5_cpu.conda
hash:
- md5: 5e467a924e4e3a48ed03fda0f3650e88
- sha256: 76638e36735c2c8f968b090814c1b583e2c8f2b56b05e155c4b82fb9f286e652
+ md5: ecde9278c41bee3cde9ddc99ee7f8d0e
+ sha256: ce351947c34f04e4b16f0d98441e02793471e043f630670d6258cb18e60a4bf5
category: main
optional: false
- name: libarrow-acero
@@ -2597,10 +1834,10 @@ package:
libarrow-compute: 26.0.1
libgcc: '>=15'
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-acero-26.0.1-h35cad7b_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-acero-26.0.1-h35cad7b_5_cpu.conda
hash:
- md5: a3e52de423eacd0f8fc45411b96e730c
- sha256: 3e895c0ddfd462b691d1d21fcd045cd59c5e34c3a348fb8c4eba05cebf27c037
+ md5: 74b436e99ef9bfda9c99592f586acf89
+ sha256: 339c364cd77e7afa72dbbcb9dc906a25c64421889bb2c334dc1c5016110bdbfd
category: main
optional: false
- name: libarrow-compute
@@ -2613,12 +1850,12 @@ package:
libgcc: '>=15'
libre2-11: '>=2025.11.5'
libstdcxx: '>=15'
- libutf8proc: '>=2.11.3,<2.12.0a0'
+ libutf8proc: '>=2.12.0,<2.13.0a0'
re2: ''
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-compute-26.0.1-h2e1e52a_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-compute-26.0.1-hc9b89cc_5_cpu.conda
hash:
- md5: f01eb652e0b2de6156317edd2fbd9d39
- sha256: 0d99a653576daad68e2bc356719ec2ea708e143c86f6b8a4fd9a9c7c4bacf9af
+ md5: 22def3abed674d30d38edfad0416f289
+ sha256: a1ca1f213dee003cee364b69c9ad596e3e01e33d87051c2be0626bd59a426f62
category: main
optional: false
- name: libarrow-dataset
@@ -2633,10 +1870,10 @@ package:
libgcc: '>=15'
libparquet: 26.0.1
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-dataset-26.0.1-h35cad7b_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-dataset-26.0.1-h35cad7b_5_cpu.conda
hash:
- md5: bf611bb0386926bc33641ac13c3b371b
- sha256: a57b26262304eb4afe3c11821f4391885a42dee58da8f733a44b058cb7c6e0fa
+ md5: 74ed7c8f31879eba79eaf6884c97aea2
+ sha256: 7d3c12fd832d10ea2f0e81461f0063420bbbf66219213eec97d227a6eec7fb6f
category: main
optional: false
- name: libarrow-s3
@@ -2650,10 +1887,10 @@ package:
libarrow: 26.0.1
libgcc: '>=15'
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-s3-26.0.1-hb32f990_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-s3-26.0.1-hb32f990_5_cpu.conda
hash:
- md5: 9d862a322d87f13aa11ce5191192c644
- sha256: 2e7a8fd5e018e8a450d988da69821759c4876ba809ec8a7ecd15a5167ddd1835
+ md5: b638106455b70e550b443bb532804902
+ sha256: 509cdde99f5b0ba5f7bd76ae7f151021ab5957ca56c39ea265db42b818dfb8d9
category: main
optional: false
- name: libarrow-substrait
@@ -2669,10 +1906,10 @@ package:
libgcc: '>=15'
libprotobuf: '>=7.35.1,<7.35.2.0a0'
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-substrait-26.0.1-hbad6200_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libarrow-substrait-26.0.1-hbad6200_5_cpu.conda
hash:
- md5: 649c9bda1221f60d347a9f5338fcd190
- sha256: d32638cb0700029202ca6e2563499a823707912add0c42d68b563bc821f94c03
+ md5: 4be0c900d1f2acb36d2954c74ec4ff02
+ sha256: 18764bdb42e52ae903e431b0f40f27ee562324174a3f2bc16ef3c0658dfba32c
category: main
optional: false
- name: libblas
@@ -2816,418 +2053,120 @@ package:
md5: f51e24ce110ae24c92074736a308e47e
sha256: 69540315b4b8de93b383243334151ed19e98968baaa59440ba645a3bff68d765
category: main
- optional: false
-- name: libclang-cpp23.1
- version: 23.1.4
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=15'
- libllvm23: '>=23.1.4,<23.2.0a0'
- libstdcxx: '>=15'
- libxml2: ''
- libxml2-16: '>=2.15.4'
- libzlib: '>=1.3.2,<2.0a0'
- zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libclang-cpp23.1-23.1.4-default_h0acdd01_1.conda
- hash:
- md5: c3831139a4d948c287b4336f1b2059fd
- sha256: 93dde2595e2c339d687be470e1f458f7641a52bda06762055db19237cbf40fac
- category: main
- optional: false
-- name: libclang13
- version: 23.1.4
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libclang-cpp23.1: ==23.1.4
- libgcc: '>=15'
- libllvm23: '>=23.1.4,<23.2.0a0'
- libstdcxx: '>=15'
- libxml2: ''
- libxml2-16: '>=2.15.4'
- libzlib: '>=1.3.2,<2.0a0'
- zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libclang13-23.1.4-default_h6e6c352_1.conda
- hash:
- md5: 2ed7cfa18d17f82b546ee2ed46bf2f7b
- sha256: 0d973a298e2e9efb1f563e0e1f2a39cc7704c4393e1aad1128c95b9b85bff1d1
- category: main
- optional: false
-- name: libcolamd
- version: 3.3.4
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=13'
- libsuitesparseconfig: '>=7.10.1,<8.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcolamd-3.3.4-hf02c80a_7100101.conda
- hash:
- md5: 56d4c5542887e8955f21f8546ad75d9d
- sha256: 00d1b976b914f0c20ae6f81f4e4713fa87717542eba8757b9a3c9e8abcc29858
- category: main
- optional: false
-- name: libcrc32c
- version: 1.1.2
- manager: conda
- platform: linux-64
- dependencies:
- libgcc-ng: '>=9.4.0'
- libstdcxx-ng: '>=9.4.0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcrc32c-1.1.2-h9c3ff4c_0.tar.bz2
- hash:
- md5: c965a5aa0d5c1c37ffc62dff36e28400
- sha256: fd1d153962764433fe6233f34a72cdeed5dcf8a883a85769e8295ce940b5b0c5
- category: main
- optional: false
-- name: libcublas
- version: 12.9.2.10
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-nvrtc: ''
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcublas-12.9.2.10-h676940d_0.conda
- hash:
- md5: 55dfb580a84aea59185586a306bf9605
- sha256: 16666002786d59dd7d62ddbeffe08c4118d882c986c6bf72443c7f983c24fb0a
- category: main
- optional: false
-- name: libcublas-dev
- version: 12.9.2.10
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-crt-dev_linux-64: ''
- cuda-cudart-dev_linux-64: ''
- cuda-version: '>=12.9,<12.10.0a0'
- libcublas: 12.9.2.10
- libgcc: '>=14'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcublas-dev-12.9.2.10-h676940d_0.conda
- hash:
- md5: 7f35376e196eecc81ffe97f63b2b7e3a
- sha256: 3a68ef147c0a7fdd59286e49c4dd1a03be9ea6180b12ee5324ecd555c17419bb
- category: main
- optional: false
-- name: libcudf
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-nvrtc: ''
- cuda-version: '>=12,<13.0a0'
- dlpack: '>=0.8,<1.0'
- libcufile: ''
- libgcc: '>=14'
- libkvikio: 26.4.*
- libnvcomp: '>=5.1.0.21,<6.0a0'
- libnvcomp-dev: 5.1.0.21.*
- libnvjitlink: '>=12.9.86,<13.0a0'
- librmm: 26.4.*
- libstdcxx: '>=14'
- rapids-logger: 0.2.*
- url: https://conda.anaconda.org/rapidsai/linux-64/libcudf-26.04.00-cuda12_260408_f9c3cf19.conda
- hash:
- md5: 9a61449338c44685176ec959a2645df2
- sha256: e2f358d6097f49b8b4f4d1f3c1832ae3c64b0a08706dfcd4bc1a273af0c6d613
- category: main
- optional: false
-- name: libcufft
- version: 11.4.1.4
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcufft-11.4.1.4-hecca717_1.conda
- hash:
- md5: 75ae571353ec92c8f34d4cf6ec6ba264
- sha256: 62d4214c182c89cfb02271a42eaac56a41f50bbbea3b0d795a8e33f167a39a4e
- category: main
- optional: false
-- name: libcufile
- version: 1.14.1.1
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- rdma-core: '>=59.0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcufile-1.14.1.1-hbc026e6_1.conda
- hash:
- md5: cab1818eada3952ed09c8dcbb7c26af7
- sha256: 5fa43e8a8d335fc0c3a6aeb2e7b0debc7d8495b8a60a56ac30f23b0e852ab74a
- category: main
- optional: false
-- name: libcufile-dev
- version: 1.14.1.1
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libcufile: 1.14.1.1
- libgcc: '>=14'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcufile-dev-1.14.1.1-hecca717_1.conda
- hash:
- md5: 74a93e4c8fd24d535abc546c6fee2155
- sha256: 3a7c726419017319df149ff67ea3efae37d668ecfc1aa2ac3273b21c4c76d68b
- category: main
- optional: false
-- name: libcugraph
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-profiler-api: ''
- cuda-version: '>=12,<13.0a0'
- libcublas: ''
- libcurand: ''
- libcusolver: ''
- libcusparse: ''
- libgcc: '>=14'
- libraft: 26.4.*
- librmm: 26.4.*
- libstdcxx: '>=14'
- nccl: '>=2.19'
- url: https://conda.anaconda.org/rapidsai/linux-64/libcugraph-26.04.00-cuda12_260408_03ba9c2d.conda
- hash:
- md5: 69673dde9391e8a217c12c8b7e5aac02
- sha256: de7d45e14dc6120a21f837ecc740517f135d46ac1200f8fe81c77d1b98a3e191
- category: main
- optional: false
-- name: libcuml
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-cudart: ''
- cuda-version: '>=12,<13.0a0'
- libcublas: ''
- libcufft: ''
- libcurand: ''
- libcusolver: ''
- libcusparse: ''
- libcuvs: 26.4.*
- libgcc: '>=14'
- librmm: 26.4.*
- libstdcxx: '>=14'
- rapids-logger: 0.2.*
- treelite: '>=4.7.0,<5.0.0'
- url: https://conda.anaconda.org/rapidsai/linux-64/libcuml-26.04.00-cuda12_260409_f223d89b.conda
- hash:
- md5: 38a2227873475e56f2d596198ac9c804
- sha256: c2b0cc9f402b44f0773fb2703209e08be92d76ac8786822c905eff558d38f478
- category: main
- optional: false
-- name: libcups
- version: 2.3.3
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- krb5: '>=1.22.2,<1.23.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- libzlib: '>=1.3.1,<2.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcups-2.3.3-h7a8fb5f_6.conda
- hash:
- md5: 49c553b47ff679a6a1e9fc80b9c5a2d4
- sha256: 205c4f19550f3647832ec44e35e6d93c8c206782bdd620c1d7cf66237580ff9c
- category: main
- optional: false
-- name: libcurand
- version: 10.3.10.19
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcurand-10.3.10.19-h676940d_1.conda
- hash:
- md5: 2a91559a9345bedf09af8b7903deb6e6
- sha256: 3d40daf956b220cc367a6306ede1e259446fb844051bcfed87c46539cc1aaf03
- category: main
- optional: false
-- name: libcurand-dev
- version: 10.3.10.19
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libcurand: 10.3.10.19
- libgcc: '>=14'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcurand-dev-10.3.10.19-h676940d_1.conda
- hash:
- md5: fc716aaff5af15b80ccbd28b3e67672c
- sha256: b506f93e7bea6d0e060f09f4bac6db3f57586084ac309db0d44b3756f5b0bc80
- category: main
- optional: false
-- name: libcurl
- version: 8.23.0
+ optional: false
+- name: libclang-cpp23.1
+ version: 23.1.5
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- krb5: '>=1.22.2,<1.23.0a0'
libgcc: '>=15'
- libnghttp2: '>=1.68.1,<2.0a0'
- libpsl: '>=0.23.1,<0.24.0a0'
- libssh2: '>=1.11.1,<2.0a0'
+ libllvm23: '>=23.1.5,<23.2.0a0'
+ libstdcxx: '>=15'
+ libxml2: ''
+ libxml2-16: '>=2.15.4'
libzlib: '>=1.3.2,<2.0a0'
- openssl: '>=3.5.9,<4.0a0'
zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcurl-8.23.0-ha042cf0_0.conda
- hash:
- md5: 0ad16151226df493aed967f0efce0e80
- sha256: d2db14d80efb4f2e76e76561c633f1767f7543987f2bcb54a067a9d264d432ab
- category: main
- optional: false
-- name: libcusolver
- version: 11.7.5.82
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libcublas: '>=12.9.1.4,<12.10.0a0'
- libcusparse: '>=12.5.10.65,<12.6.0a0'
- libgcc: '>=14'
- libnvjitlink: '>=12.9.86,<13.0a0'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcusolver-11.7.5.82-h676940d_2.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libclang-cpp23.1-23.1.5-default_h0acdd01_0.conda
hash:
- md5: bb6e31a0daa64ede76fe8d3fff01c06f
- sha256: 8691cf6b1585cf6251663029e00485da5a912f6ca0ff7e5c31a6d8d604b29253
+ md5: 14c1b208c11eea04cef332636b002394
+ sha256: 0c506f0e598c90e6fe374188e6e102a1346dcbf2b43d25e3db3e3975700d8a5d
category: main
optional: false
-- name: libcusolver-dev
- version: 11.7.5.82
+- name: libclang13
+ version: 23.1.5
manager: conda
platform: linux-64
dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libcusolver: 11.7.5.82
- libgcc: '>=14'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcusolver-dev-11.7.5.82-h676940d_2.conda
+ __glibc: '>=2.17,<3.0.a0'
+ libclang-cpp23.1: ==23.1.5
+ libgcc: '>=15'
+ libllvm23: '>=23.1.5,<23.2.0a0'
+ libstdcxx: '>=15'
+ libxml2: ''
+ libxml2-16: '>=2.15.4'
+ libzlib: '>=1.3.2,<2.0a0'
+ zstd: '>=1.5.7,<1.6.0a0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libclang13-23.1.5-default_h782a7a8_0.conda
hash:
- md5: 0fe12e558abf507458bcec839e29778d
- sha256: f2a974af90ecf6e47c2780741b5351c5f21d20bf6b9fb4448966f07d23ad27b8
+ md5: 77a10c771b02cfd2f433d43b64298892
+ sha256: 727fb7f72d8f9aedea7a00e0d043bbc87153c9aa6951dd21e446e75378053776
category: main
optional: false
-- name: libcusparse
- version: 12.5.10.65
+- name: libcolamd
+ version: 3.3.4
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libgcc: '>=14'
- libnvjitlink: '>=12.9.86,<13.0a0'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcusparse-12.5.10.65-hecca717_2.conda
+ libgcc: '>=13'
+ libsuitesparseconfig: '>=7.10.1,<8.0a0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libcolamd-3.3.4-hf02c80a_7100101.conda
hash:
- md5: 890ebfaad48c887d3d82847ec9d6bc79
- sha256: 7b511549a22df408d36dadbeabdfd9c35b124d9d6f000b29ffcbe4b38b7faeb7
+ md5: 56d4c5542887e8955f21f8546ad75d9d
+ sha256: 00d1b976b914f0c20ae6f81f4e4713fa87717542eba8757b9a3c9e8abcc29858
category: main
optional: false
-- name: libcusparse-dev
- version: 12.5.10.65
+- name: libcrc32c
+ version: 1.1.2
manager: conda
platform: linux-64
dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-version: '>=12.9,<12.10.0a0'
- libcusparse: 12.5.10.65
- libgcc: '>=14'
- libnvjitlink: '>=12.9.86,<13.0a0'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libcusparse-dev-12.5.10.65-hecca717_2.conda
+ libgcc-ng: '>=9.4.0'
+ libstdcxx-ng: '>=9.4.0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libcrc32c-1.1.2-h9c3ff4c_0.tar.bz2
hash:
- md5: db94469fbd554c107acc3afd0af5d8ec
- sha256: c6d7ec3ccef6dce988c3acc93198973ec9ff5aa9ffe99e07dd953c2d3b409a3b
+ md5: c965a5aa0d5c1c37ffc62dff36e28400
+ sha256: fd1d153962764433fe6233f34a72cdeed5dcf8a883a85769e8295ce940b5b0c5
category: main
optional: false
-- name: libcuvs
- version: 26.04.00
+- name: libcups
+ version: 2.3.3
manager: conda
platform: linux-64
dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-cudart: ''
- cuda-version: '>=12,<13.0a0'
- libcublas: ''
- libcurand: ''
- libcusolver: ''
- libcusparse: ''
- libcuvs-headers: ==26.4.0
+ __glibc: '>=2.17,<3.0.a0'
+ krb5: '>=1.22.2,<1.23.0a0'
libgcc: '>=14'
- libraft-headers: 26.4.*
- librmm: 26.4.*
libstdcxx: '>=14'
- nccl: ''
- url: https://conda.anaconda.org/rapidsai/linux-64/libcuvs-26.04.00-cuda12_260408_60b19a0f.conda
+ libzlib: '>=1.3.1,<2.0a0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libcups-2.3.3-h7a8fb5f_6.conda
hash:
- md5: 2b9d59fc5a9ee033d42f68ef6cd451aa
- sha256: b7f4aee61e7c0e576d847664d894d2e60bb49b564767440350e0e050f6316da7
+ md5: 49c553b47ff679a6a1e9fc80b9c5a2d4
+ sha256: 205c4f19550f3647832ec44e35e6d93c8c206782bdd620c1d7cf66237580ff9c
category: main
optional: false
-- name: libcuvs-headers
- version: 26.04.00
+- name: libcurl
+ version: 8.23.0
manager: conda
platform: linux-64
dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-cudart: ''
- cuda-version: '>=12,<13.0a0'
- libcublas: ''
- libcurand: ''
- libcusolver: ''
- libcusparse: ''
- libgcc: '>=14'
- libraft-headers: 26.4.*
- librmm: 26.4.*
- libstdcxx: '>=14'
- nccl: ''
- url: https://conda.anaconda.org/rapidsai/linux-64/libcuvs-headers-26.04.00-cuda12_260408_60b19a0f.conda
+ __glibc: '>=2.17,<3.0.a0'
+ krb5: '>=1.22.2,<1.23.0a0'
+ libgcc: '>=15'
+ libnghttp2: '>=1.68.1,<2.0a0'
+ libpsl: '>=0.23.1,<0.24.0a0'
+ libssh2: '>=1.11.1,<2.0a0'
+ libzlib: '>=1.3.2,<2.0a0'
+ openssl: '>=3.5.9,<4.0a0'
+ zstd: '>=1.5.7,<1.6.0a0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libcurl-8.23.0-ha042cf0_0.conda
hash:
- md5: c3216f4b246d2016b1f1a9d719dc94ad
- sha256: 491be453964808a7b527772f30e247625cc3c5add40683263d2abb349793b886
+ md5: 0ad16151226df493aed967f0efce0e80
+ sha256: d2db14d80efb4f2e76e76561c633f1767f7543987f2bcb54a067a9d264d432ab
category: main
optional: false
- name: libdeflate
- version: '1.25'
+ version: '1.26'
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libgcc: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libdeflate-1.25-hd45a770_1.conda
+ libgcc: '>=15'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libdeflate-1.26-h351257e_0.conda
hash:
- md5: 40f9b31aa9cf007789867df0decd0492
- sha256: 82e134c8a08b1eed9a2ed8ab578b89aa1730dcde3dea8dd87645ed0637878e54
+ md5: d0c17fc12939a990f0865b7a14b49cfe
+ sha256: f437f68abb1fa1dededd082a0c2800bec40495e01cf9ec46f4e693e34d60b644
category: main
optional: false
- name: libdrm
@@ -3300,17 +2239,17 @@ package:
category: main
optional: false
- name: libevent
- version: 2.1.12
+ version: 2.1.13
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
libgcc: '>=15'
- openssl: '>=3.5.8,<4.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libevent-2.1.12-h5348a74_2.conda
+ openssl: '>=3.5.9,<4.0a0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libevent-2.1.13-h5348a74_0.conda
hash:
- md5: 2e699c6e151045b2cfea34b53ed69439
- sha256: f874779772878f3b509c607de9089c282f71232fde4cac876fc660d29670f28d
+ md5: 53b60cc0d314d0f1ebfdbbf7d8f6e313
+ sha256: 57ed3320311043c9bc829825fa6e30db3e17307acdd8f5b7bbc917bb36247bf4
category: main
optional: false
- name: libexpat
@@ -3472,7 +2411,7 @@ package:
category: main
optional: false
- name: libglib
- version: 2.90.0
+ version: 2.90.1
manager: conda
platform: linux-64
dependencies:
@@ -3482,10 +2421,10 @@ package:
libiconv: '>=1.18,<2.0a0'
libzlib: '>=1.3.2,<2.0a0'
pcre2: '>=10.47,<10.48.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libglib-2.90.0-h569388d_0.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libglib-2.90.1-h569388d_1.conda
hash:
- md5: c85e9539551f1fd21104e518d940c244
- sha256: 9e688e046550415c030c3d08ed9a3fd13a51d7bc56295283b858b6c3070bb9f8
+ md5: 2b4731a8f8cd069c1d177f972e9d0d99
+ sha256: 463a24904fd83a3223c9f292211975bba405838ff32e6179df04a3cca202a2b3
category: main
optional: false
- name: libglvnd
@@ -3681,24 +2620,6 @@ package:
sha256: 6b4d462642c240dc3671af74f7705b23f34eea0f71e0d9dbcf14b4ed008311ff
category: main
optional: false
-- name: libkvikio
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-version: '>=12.2.0a0,<13.0a0'
- libcufile-dev: ''
- libcurl: '>=8.13.0,<9.0a0'
- libgcc: '>=14,>=15'
- libnuma: '>=2.0.18,<3.0a0'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/rapidsai/linux-64/libkvikio-26.04.00-cuda12_260408_5aba1d10.conda
- hash:
- md5: 1a64a7d741811c880686e3265d01414a
- sha256: f932555ff69ec7fdea7349c8a92cb5502eccf4ecd3f9dacb1c57e621a0e1d535
- category: main
- optional: false
- name: liblapack
version: 3.11.0
manager: conda
@@ -3829,90 +2750,6 @@ package:
sha256: 3b3f19ced060013c2dd99d9d46403be6d319d4601814c772a3472fe2955612b0
category: main
optional: false
-- name: libnuma
- version: 2.0.18
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libnuma-2.0.18-hb03c661_4.conda
- hash:
- md5: bfe9cd9c6d92170947301fef63356e39
- sha256: db0a0fc67104196afe6732ff055b30829f984f4cf1e8efff4ea5b2bbb1640da2
- category: main
- optional: false
-- name: libnvcomp
- version: 5.1.0.21
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-version: '>=12,<13.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libnvcomp-5.1.0.21-hb7e823c_0.conda
- hash:
- md5: c3b09b2378d35bbc72bbbe2a47c11240
- sha256: 353bc179da18fa1d3c7e02120f248719356b622e8af51456c6d59c11d87f4922
- category: main
- optional: false
-- name: libnvcomp-dev
- version: 5.1.0.21
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-version: '>=12,<13.0a0'
- libgcc: '>=14'
- libnvcomp: 5.1.0.21
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libnvcomp-dev-5.1.0.21-hb7e823c_0.conda
- hash:
- md5: 3dfd1f10e4229815cd96450ca20b45d6
- sha256: 0afc084708324ae72eaa27defd43107b19695150394fbdf90b40735125153e08
- category: main
- optional: false
-- name: libnvjitlink
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-version: '>=12,<12.10.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libnvjitlink-12.9.86-hecca717_2.conda
- hash:
- md5: 3461b0f2d5cbb7973d361f9e85241d98
- sha256: 3b1c851f4fc42d347ce1c1606bdd195343a47f121e0fceb7a1f1e5aa1d497da9
- category: main
- optional: false
-- name: libnvptxcompiler-dev
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- cuda-version: '>=12.9,<12.10.0a0'
- libnvptxcompiler-dev_linux-64: 12.9.86
- url: https://conda.anaconda.org/conda-forge/linux-64/libnvptxcompiler-dev-12.9.86-ha770c72_2.conda
- hash:
- md5: 3fd926c321c6dbf386aa14bd8b125bfb
- sha256: 1e7a7b34f8639a5feb75ba864127059e4d83edfe1a516547f0dbb9941e7b8f8b
- category: main
- optional: false
-- name: libnvptxcompiler-dev_linux-64
- version: 12.9.86
- manager: conda
- platform: linux-64
- dependencies:
- cuda-version: '>=12.9,<12.10.0a0'
- url: https://conda.anaconda.org/conda-forge/noarch/libnvptxcompiler-dev_linux-64-12.9.86-ha770c72_2.conda
- hash:
- md5: a66a909acf08924aced622903832a937
- sha256: 17952c32eac197a59c119fdf3fb6f08c6a29c225a80bae141ac904ad212b87dd
- category: main
- optional: false
- name: libopenblas
version: 0.3.34
manager: conda
@@ -3985,10 +2822,10 @@ package:
libthrift: '>=0.25.0,<0.25.1.0a0'
openssl: '>=3.5.9,<4.0a0'
simdjson: '>=5.1.1,<5.2.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libparquet-26.0.1-h2d6894f_2_cpu.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libparquet-26.0.1-h2d6894f_5_cpu.conda
hash:
- md5: d323682fdf093fec93a918e1c3b4d7a8
- sha256: 17b77aa533b3496c075b2428743c41a3e1baee2485fd6f85583402664e5d2ff8
+ md5: fe532df58c640bd0eb6215bc45a3e4ce
+ sha256: 96d1d46a7b7d4df43a49b8555195965b8587075df1b5d88adc7b7bcc326fe4a8
category: main
optional: false
- name: libpciaccess
@@ -4029,11 +2866,11 @@ package:
libgcc: '>=15'
libstdcxx: '>=15'
openldap: '>=2.6.13,<2.7.0a0'
- openssl: '>=3.5.8,<4.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libpq-18.6-h3883810_3.conda
+ openssl: '>=3.5.9,<4.0a0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libpq-18.6-h3883810_4.conda
hash:
- md5: cc998047391971a452fed680158329f3
- sha256: 60fa3b973f6f68a367171336382d801404ee7b38c987c3b6a55c207086f736de
+ md5: 13266bfcec35113306ccb14fb578dfae
+ sha256: 94b9a4abf3bd67bb772d8050ae3933b7987b0a9a28ca354b35c45c2ebd93b796
category: main
optional: false
- name: libprotobuf
@@ -4093,59 +2930,6 @@ package:
sha256: 71126cad347dfd1be645b1207baa3729e64f055edb5f785dadeb84872480a015
category: main
optional: false
-- name: libraft
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- cuda-profiler-api: ''
- cuda-version: '>=12,<13.0a0'
- libcublas: ''
- libcurand: ''
- libcusolver: ''
- libcusparse: ''
- libraft-headers: ==26.4.0
- url: https://conda.anaconda.org/rapidsai/linux-64/libraft-26.04.00-cuda12_260408_b01e3028.conda
- hash:
- md5: 04014b08138ccd053a8352d879f1cb1e
- sha256: 449236d3f94323271d540add3daf978195784723b24178038c537e2f2efcf8cf
- category: main
- optional: false
-- name: libraft-headers
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- cuda-profiler-api: ''
- cuda-version: '>=12,<13.0a0'
- libcublas-dev: ''
- libcurand-dev: ''
- libcusolver-dev: ''
- libcusparse-dev: ''
- libraft-headers-only: ==26.4.0
- librmm: 26.4.*
- nccl: '>=2.19'
- ucxx: 0.49.*
- url: https://conda.anaconda.org/rapidsai/linux-64/libraft-headers-26.04.00-cuda12_260408_b01e3028.conda
- hash:
- md5: 413e57142e096269834e8a95b4905d09
- sha256: 0913ebc5a541d2d8f695d14245008470cd05232badd7df23524db6eeaae0fd56
- category: main
- optional: false
-- name: libraft-headers-only
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- cuda-version: '>=12,<13.0a0'
- librmm: 26.4.*
- rapids-logger: 0.2.*
- url: https://conda.anaconda.org/rapidsai/linux-64/libraft-headers-only-26.04.00-cuda12_260408_b01e3028.conda
- hash:
- md5: 4c79a5c4309aa770d00749683f941c70
- sha256: 9c062e000afecbc9cdd6da27da3b32a7691d77a71b769af25b80fbfa89f5ea00
- category: main
- optional: false
- name: libraqm
version: 0.11.0
manager: conda
@@ -4172,27 +2956,10 @@ package:
libabseil: '*,>=20260817.0,<20260818.0a0'
libgcc: '>=15'
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/libre2-11-2025.11.05-h02ca46f_3.conda
- hash:
- md5: ad950bb7d6c2b0a2cb62d9af66c9a082
- sha256: 8310dbf66d801a577349eb6268a6ef83accc5f5a042b2911a7d7273a9f991105
- category: main
- optional: false
-- name: librmm
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-cudart: ''
- cuda-version: '>=12,<13.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- rapids-logger: 0.2.*
- url: https://conda.anaconda.org/rapidsai/linux-64/librmm-26.04.00-cuda12_260408_48b36cc6.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libre2-11-2025.11.05-h02ca46f_4.conda
hash:
- md5: 815e224bae37f1bf6491fe3145adbda2
- sha256: 3b98126a0e3407c7291580a41adf758af7844f4ecae3578de23a895f2ecb3872
+ md5: 2992bcf0ff620215b8427fcafad4afe8
+ sha256: 0c49508a9525f89e9195155e106ae8c9fb57ddf46e5111f98d7f3227976c6a15
category: main
optional: false
- name: libscotch
@@ -4334,15 +3101,15 @@ package:
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libevent: '>=2.1.12,<2.1.13.0a0'
+ libevent: '>=2.1.13,<2.1.14.0a0'
libgcc: '>=15'
libstdcxx: '>=15'
libzlib: '>=1.3.2,<2.0a0'
openssl: '>=3.5.9,<4.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libthrift-0.25.0-h22fcfe2_1.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libthrift-0.25.0-h06f262d_2.conda
hash:
- md5: 623422483280b48e40139e11f7991f31
- sha256: 70c6e9d07c960739ebbac5519225aa0f80556ad1755597c99f7188eeeb37fa33
+ md5: 3a01b39543e4b3e1e8dcb74f0935727a
+ sha256: e3ea13e90078bd3a87162a2b4c030507f0d0bdad1a6064f9ab5e19321b944f45
category: main
optional: false
- name: libtiff
@@ -4351,36 +3118,19 @@ package:
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- lerc: '>=4.2.0,<5.0a0'
- libdeflate: '>=1.25,<1.26.0a0'
+ lerc: '>=4.3.0,<5.0a0'
+ libdeflate: '>=1.26,<1.27.0a0'
libgcc: '>=15'
- libjpeg-turbo: '>=3.2.0,<4.0a0'
+ libjpeg-turbo: '>=3.2.1,<4.0a0'
liblzma: '>=5.8.3,<6.0a0'
libstdcxx: '>=15'
libwebp-base: '>=1.6.0,<2.0a0'
libzlib: '>=1.3.2,<2.0a0'
zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libtiff-4.7.2-hcc2c06a_1.conda
- hash:
- md5: 0907d876f460ebc941ae590338b6102f
- sha256: 10e125da82ca93e09191e66e5a94d566b2cf8d4024e2a0db3a9114297447e0d9
- category: main
- optional: false
-- name: libucxx
- version: 0.49.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-version: '>=12,<13.0a0'
- libgcc: '>=14'
- librmm: '>=26.4.0,<26.5.0a0'
- libstdcxx: '>=14'
- ucx: '>=1.18.0,<1.21.0'
- url: https://conda.anaconda.org/rapidsai/linux-64/libucxx-0.49.00-cuda12_260408_8d47a9ff.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libtiff-4.7.2-h6d37cbf_2.conda
hash:
- md5: b8ddfc6e13aa7b755392fe8cb6f9250a
- sha256: f1153dd5eb1492c2ae820bc4a5eaa8ede66043385c598540f2b4884d4d8eef9d
+ md5: 18023fe8fbec4f031ca88444572590e4
+ sha256: 3406eed96ee929309ecd3937308d5a81302f4ca2eddd30e832071cbe9121a6c7
category: main
optional: false
- name: libudev1
@@ -4415,16 +3165,16 @@ package:
category: main
optional: false
- name: libutf8proc
- version: 2.11.3
+ version: 2.12.0
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libgcc: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/libutf8proc-2.11.3-hfe17d71_0.conda
+ libgcc: '>=15'
+ url: https://conda.anaconda.org/conda-forge/linux-64/libutf8proc-2.12.0-hc4fb7df_0.conda
hash:
- md5: 1247168fe4a0b8912e3336bccdbf98a5
- sha256: ecbf4b7520296ed580498dc66a72508b8a79da5126e1d6dc650a7087171288f9
+ md5: a7d99b24a07fda1a6442ad64b759bea5
+ sha256: 0c86e7b7fd600b05ac84cf1bc049e3a562e3bdd5c37433386a611060b7856fb9
category: main
optional: false
- name: libuuid
@@ -4479,10 +3229,10 @@ package:
pthread-stubs: ''
xorg-libxau: '>=1.0.12,<2.0a0'
xorg-libxdmcp: '>=1.1.5,<2.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/libxcb-1.17.0-hb83e432_2.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/libxcb-1.17.0-hb83e432_3.conda
hash:
- md5: 61f0ffba9161cbb3a77722869b21e4bb
- sha256: 7b49e6fd2a584b7f072943e6adf4149677af5121246975278804782d37752837
+ md5: 661d9e1392a1b672d19d76df3a1feb07
+ sha256: 38021f35547021762c04860d1c9f2d3186b83e090e5f1773216045de06cf10cc
category: main
optional: false
- name: libxcrypt
@@ -4626,33 +3376,21 @@ package:
category: main
optional: false
- name: llvmlite
- version: 0.46.0
+ version: 0.50.0
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- libzlib: '>=1.3.1,<2.0a0'
- python: '>=3.12,<3.13.0a0'
+ libgcc: '>=15'
+ libstdcxx: '>=15'
+ libzlib: '>=1.3.2,<2.0a0'
+ python: ''
python_abi: 3.12.*
zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/llvmlite-0.46.0-py312h7424e68_0.conda
- hash:
- md5: 7b8f200683fab3c020c37254debfcbc5
- sha256: 1dbcff26480ae7a7a466b45aaa06b793ad66fe2a167ca2b5805e449b0403e3c0
- category: main
- optional: false
-- name: locket
- version: 1.0.0
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=2.7,!=3.0.*,!=3.1.*,!=3.2.*,!=3.3.*'
- url: https://conda.anaconda.org/conda-forge/noarch/locket-1.0.0-pyhd8ed1ab_0.tar.bz2
+ url: https://conda.anaconda.org/conda-forge/linux-64/llvmlite-0.50.0-py312hb491018_0.conda
hash:
- md5: 91e27ef3d05cc772ce627e51cff111c4
- sha256: 9afe0b5cfa418e8bdb30d8917c5a6cec10372b037924916f1f85b9f4899a67a6
+ md5: afdb20a709a3fb3f5c14d5849d11e67c
+ sha256: fbf41bdb6920ad8c0912bb5571e1b3bd087f9e1453beed7110c34a80dc40b69b
category: main
optional: false
- name: louvain
@@ -4674,22 +3412,6 @@ package:
sha256: 8dece101639c6defb94fbfcf805e924ef7c3728a4bb1b7751f2c31e460fa77d1
category: main
optional: false
-- name: lz4
- version: 4.4.5
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=15'
- lz4-c: '>=1.10.0,<1.11.0a0'
- python: ''
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/lz4-4.4.5-py312hc6a72db_2.conda
- hash:
- md5: a98e678cb247b751c1148fdc9deb5189
- sha256: 3e840f3334eb0b73056c037d7e885c84dc6f1dd77d7130ac51fd51aa1fb33d88
- category: main
- optional: false
- name: lz4-c
version: 1.10.0
manager: conda
@@ -4870,17 +3592,17 @@ package:
__glibc: '>=2.17,<3.0.a0'
libfabric: ''
libfabric1: '>=1.14.0'
- libgcc: '>=14'
+ libgcc: '>=15'
libgfortran: ''
- libgfortran5: '>=14.3.0'
+ libgfortran5: '>=15.3.0'
libhwloc: '>=2.13.0,<2.13.1.0a0'
- libstdcxx: '>=14'
+ libstdcxx: '>=15'
mpi: 1.0.*
- ucx: '>=1.20.1,<1.21.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/mpich-5.0.1-h6f9170e_1.conda
+ ucx: '>=1.22.0,<1.23.0a0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/mpich-5.0.1-hfbb40bb_4.conda
hash:
- md5: 59d4da8d56516878beaac8577d3cbd74
- sha256: 326e36d06746dd96350dbb232f9fbb0e948e71f04e4225c2d76561153aff46f0
+ md5: 5458b0034085c2c78865f67667b9d189
+ sha256: 63cd7fb6a60ed5851e070f864a5ec5109ba23dedbfe04e785fab8aff9fd7031a
category: main
optional: false
- name: msgpack-python
@@ -4914,21 +3636,6 @@ package:
sha256: 7fb73bbf9c68cf2e62f4c9625f1aa04521d55760afb3d517aad29fb51ca503c6
category: main
optional: false
-- name: multidict
- version: 7.2.1
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=15'
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/multidict-7.2.1-py312hfd18d64_0.conda
- hash:
- md5: 9811577aef8b6c9e9fffd8c551da2a02
- sha256: d14b193ec98dac05fde3feec813b28b9e138f345844968328dfcf678c9bb06a7
- category: main
- optional: false
- name: mumps-include
version: 5.8.2
manager: conda
@@ -5001,21 +3708,6 @@ package:
sha256: aeb1548eb72e4f198e72f19d242fb695b35add2ac7b2c00e0d83687052867680
category: main
optional: false
-- name: nccl
- version: 2.32.3.1
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- cuda-version: '>=12,<13.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- url: https://conda.anaconda.org/conda-forge/linux-64/nccl-2.32.3.1-h4d09622_0.conda
- hash:
- md5: ca0d8a2217b44a9c92fabc8c551cdaa8
- sha256: 256911abd99eb8a9b2be10911e77d32e7633fa33135c7471d7867fc40071be50
- category: main
- optional: false
- name: ncurses
version: '6.6'
manager: conda
@@ -5053,47 +3745,22 @@ package:
category: main
optional: false
- name: numba
- version: 0.64.0
+ version: 0.68.1
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
_openmp_mutex: '>=4.5'
- libgcc: '>=14'
- libstdcxx: '>=14'
- llvmlite: '>=0.46.0,<0.47.0a0'
- numpy: '>=1.22.3,<2.5,>=1.23,<3'
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/numba-0.64.0-py312hd1dde6f_0.conda
- hash:
- md5: cd04438ee6bd0adad4cd49f963a84ef2
- sha256: 82d6a7b5b475d8b90ea90a11ae3459424fa8924cef363a2777d7ab4a7ab9e6aa
- category: main
- optional: false
-- name: numba-cuda
- version: 0.28.2
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- cuda-bindings: '>=12.9.1,<14.0.0'
- cuda-core: '>=0.5.1,<1.0.0'
- cuda-cudart-dev: ''
- cuda-nvcc-impl: ''
- cuda-pathfinder: '>=1.3.4,<2.0.0'
- cuda-version: '>=12'
- libgcc: '>=14'
- libstdcxx: '>=14'
- numba: '>=0.60.0'
- numpy: '>=1.22,>=1.23,<3'
- packaging: ''
+ libgcc: '>=15'
+ libstdcxx: '>=15'
+ llvmlite: '>=0.50.0,<0.51.0a0'
+ numpy: '>=1.22.3,<2.6,>=1.25,<3'
python: ''
python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/numba-cuda-0.28.2-py312h14c7f5c_0.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/numba-0.68.1-py312h80505a6_0.conda
hash:
- md5: 0f105d9570578c65ffafd44016d783b2
- sha256: 8cca425920e38f089251606bf46d03e5d4025f48c154820e34e81cff0e44bcdb
+ md5: 596938cf8369321a6213a1ddefa89d59
+ sha256: 52de9882e497859ccd3b25af7bf6df98cc3a0285f047d741ebbf33494921172d
category: main
optional: false
- name: numcodecs
@@ -5115,7 +3782,7 @@ package:
category: main
optional: false
- name: numpy
- version: 2.4.6
+ version: 2.3.5
manager: conda
platform: linux-64
dependencies:
@@ -5127,37 +3794,10 @@ package:
libstdcxx: '>=14'
python: ''
python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/numpy-2.4.6-py312h33ff503_0.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/numpy-2.3.5-py312h33ff503_1.conda
hash:
- md5: 6e31d55ee1110fda83b4f4045f4d73ff
- sha256: dfcbeadb3e7ad0da7a55a0525884ca34c19584154e13cc4159396b305d1bd445
- category: main
- optional: false
-- name: nvidia-ml-py
- version: 13.615.71
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.11'
- url: https://conda.anaconda.org/conda-forge/noarch/nvidia-ml-py-13.615.71-pyhd8ed1ab_0.conda
- hash:
- md5: 5ba2c050eadf675f6f788bf95edae4fa
- sha256: d8f75aaed9935cd378c0694f8faa0af8c6c159f950daf56dfffe773d2076a6cc
- category: main
- optional: false
-- name: nvtx
- version: 0.2.17
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=15'
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/nvtx-0.2.17-py312h5cc1888_0.conda
- hash:
- md5: 38f0516bcf1f93ca0d2730b4e6641584
- sha256: 883d209d99615c895b39f6647baca5a675bca35edbb901203049cd122ff46639
+ md5: 03baecffb72fa96fe234fd505908065f
+ sha256: 76ad6a6f4761084b074a587fe1512956891f04b5250cec0fd39aca0f39ad122b
category: main
optional: false
- name: omegaconf
@@ -5250,13 +3890,13 @@ package:
libstdcxx: '>=15'
libzlib: '>=1.3.2,<2.0a0'
lz4-c: '>=1.10.0,<1.11.0a0'
- snappy: '>=1.2.2,<1.3.0a0'
+ snappy: '>=1.3.1,<1.4.0a0'
tzdata: ''
zstd: '>=1.5.7,<1.6.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/orc-2.3.1-hdc56c17_1.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/orc-2.3.1-h7b2d648_2.conda
hash:
- md5: d65dbc4db24e05c2f1c9664fb34e23d1
- sha256: 7e0edd79df3c6c70fa46e4322cd92c107cd617e3b3afc8d809b561ca105ad77d
+ md5: 82f0a2f57530e0bbfbc83aa46e1d851a
+ sha256: fb29474cac4c60dbc929e539c391303b0f47fb846459e36edc44ae841b9858f3
category: main
optional: false
- name: packaging
@@ -5272,23 +3912,21 @@ package:
category: main
optional: false
- name: pandas
- version: 2.3.3
+ version: 3.1.1
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- numpy: '>=1.22.4,>=1.23,<3'
- python: '>=3.12,<3.13.0a0'
- python-dateutil: '>=2.8.2'
- python-tzdata: '>=2022.7'
+ libgcc: '>=15'
+ libstdcxx: '>=15'
+ numpy: '>=1.25,<3,>=2.0.2'
+ python: ''
+ python-dateutil: '>=2.9.0'
python_abi: 3.12.*
- pytz: '>=2020.1'
- url: https://conda.anaconda.org/conda-forge/linux-64/pandas-2.3.3-py312hf79963d_1.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/pandas-3.1.1-np2py312h91ec553_0.conda
hash:
- md5: e597b3e812d9613f659b7d87ad252d18
- sha256: f633d5f9b28e4a8f66a6ec9c89ef1b6743b880b0511330184b4ab9b7e2dda247
+ md5: d3c0f076ef954125c9c3db84794bc51b
+ sha256: bcbf9b1c9c040b343791536fdd6c5acd87e73f0c2d58859c9164756528626b34
category: main
optional: false
- name: parmetis
@@ -5318,20 +3956,6 @@ package:
sha256: 611882f7944b467281c46644ffde6c5145d1a7730388bcde26e7e86819b0998e
category: main
optional: false
-- name: partd
- version: 1.4.2
- manager: conda
- platform: linux-64
- dependencies:
- locket: ''
- python: '>=3.9'
- toolz: ''
- url: https://conda.anaconda.org/conda-forge/noarch/partd-1.4.2-pyhd8ed1ab_0.conda
- hash:
- md5: 0badf9c54e24cecfb0ad2f99d680c163
- sha256: 472fc587c63ec4f6eba0cc0b06008a6371e0a08a5986de3cf4e8024a47b4fe6c
- category: main
- optional: false
- name: patsy
version: 1.0.3
manager: conda
@@ -5512,6 +4136,21 @@ package:
sha256: 8bbb25f131001457e5fd19385000e9d6a8047223a35f4c507d92113eb68d42f9
category: main
optional: false
+- name: pooch
+ version: 1.9.0
+ manager: conda
+ platform: linux-64
+ dependencies:
+ packaging: '>=20.0'
+ platformdirs: '>=2.5.0'
+ python: '>=3.10'
+ requests: '>=2.19.0'
+ url: https://conda.anaconda.org/conda-forge/noarch/pooch-1.9.0-pyhd8ed1ab_0.conda
+ hash:
+ md5: dd4b6337bf8886855db6905b336db3c8
+ sha256: 081e52c4612830bf1fd4a9c78eebaf335d1385d74ddfd328b1b2f26b983848eb
+ category: main
+ optional: false
- name: prometheus-cpp
version: 1.3.0
manager: conda
@@ -5542,21 +4181,6 @@ package:
sha256: efe8def2c93aa34cd8d3c9af1dc4c7d312791cf769d8b2b615e32733e6df6051
category: main
optional: false
-- name: propcache
- version: 0.5.4
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=15'
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/propcache-0.5.4-py312hfd18d64_2.conda
- hash:
- md5: 30ef631f5a7e1840b51937fb502e89db
- sha256: e045edbe54d9b7301b981a92057b73bc205200d0989a09542a93c38b89a43126
- category: main
- optional: false
- name: psutil
version: 7.2.2
manager: conda
@@ -5723,71 +4347,6 @@ package:
sha256: f5f015ff1bc3e1b7fc08eee096b1865234189ae0b82bebdb86a5369116e42fa0
category: main
optional: false
-- name: pylibcudf
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- _python_abi3_support: 1.*
- cpython: '>=3.11'
- cuda-python: '>=12.9.2,<13.0'
- cuda-version: '>=12,<13.0a0'
- libcudf: 26.4.0.*
- libgcc: '>=14'
- libstdcxx: '>=14'
- nvtx: '>=0.2.1'
- python: ''
- rmm: '>=26.4.0,<26.5.0a0'
- url: https://conda.anaconda.org/rapidsai/linux-64/pylibcudf-26.04.00-cuda12_cp311_abi3_260408_f9c3cf19.conda
- hash:
- md5: e09c078daeb8adc969cd0aa853101d4c
- sha256: 29175d6f55a50d591fd4ff21fbdc3a20cab251660a0825c4cb81df02f1fa9ab5
- category: main
- optional: false
-- name: pylibcugraph
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- _python_abi3_support: 1.*
- cpython: '>=3.11'
- cuda-version: '>=12,<13.0a0'
- cupy: '>=13.6.0'
- libcugraph: 26.4.0.*
- libgcc: '>=14'
- libstdcxx: '>=14'
- numpy: '>=1.23,<3.0'
- pylibraft: 26.4.*
- python: ''
- rmm: 26.4.*
- url: https://conda.anaconda.org/rapidsai/linux-64/pylibcugraph-26.04.00-cuda12_cp311_abi3_260408_03ba9c2d.conda
- hash:
- md5: fcb0bd0b040f1dd12dcfb59ea8099c92
- sha256: 7f75cdf7c22863b74b8347b86bb5e4a7edc4564df17975ccd1f7fe6bec8c5659
- category: main
- optional: false
-- name: pylibraft
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- _python_abi3_support: 1.*
- cpython: '>=3.11'
- cuda-python: '>=12.9.2,<13.0'
- cuda-version: '>=12,<13.0a0'
- libraft: 26.4.0.*
- libraft-headers: 26.4.0.*
- numpy: '>=1.23,<3.0'
- python: ''
- rmm: 26.4.*
- url: https://conda.anaconda.org/rapidsai/linux-64/pylibraft-26.04.00-cuda12_cp311_abi3_260408_b01e3028.conda
- hash:
- md5: 45e40019a250657411b7b42182c33dfd
- sha256: 34fe31f5fd0a000fb5de95de91097e335daafe92ff5845940f09f80ec05c645d
- category: main
- optional: false
- name: pynndescent
version: 0.7.0
manager: conda
@@ -5995,18 +4554,6 @@ package:
sha256: ee9c2922e07afc85fc82d5fa82c9ac2c79da3be3283a5e17bf2f2ae38dbd02fb
category: main
optional: false
-- name: pytz
- version: '2026.6'
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.11'
- url: https://conda.anaconda.org/conda-forge/noarch/pytz-2026.6-pyh5ded981_0.conda
- hash:
- md5: 9b60b8aadafa78f7581c49a057764e7f
- sha256: 624d4243a9a2ebec6349587d34bd64c8b04ab6ed174acf102e02a441afde5ae0
- category: main
- optional: false
- name: pyyaml
version: 6.0.3
manager: conda
@@ -6103,56 +4650,6 @@ package:
sha256: d3d45dd34724226eb54065caa77ae166fbb971a5dfdff5e5c66e7c03b6b5157c
category: main
optional: false
-- name: raft-dask
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- _python_abi3_support: 1.*
- cpython: '>=3.11'
- cuda-python: '>=12.9.2,<13.0'
- cuda-version: '>=12,<13.0a0'
- dask-cuda: 26.4.*
- distributed-ucxx: 0.49.*
- nccl: '>=2.19'
- pylibraft: 26.4.0.*
- python: ''
- rapids-dask-dependency: 26.4.*
- rmm: 26.4.*
- url: https://conda.anaconda.org/rapidsai/linux-64/raft-dask-26.04.00-cuda12_cp311_abi3_260408_b01e3028.conda
- hash:
- md5: 2cbbcdb299c9acc8dda901f351b66526
- sha256: e73179fe8badc9e23a48d1a64e4241981644d4305b0d259c1f3ae0c4f97ae81c
- category: main
- optional: false
-- name: rapids-dask-dependency
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- dask: 2026.1.1
- dask-core: 2026.1.1
- distributed: 2026.1.1
- url: https://conda.anaconda.org/rapidsai/noarch/rapids-dask-dependency-26.04.00-260408_8071384e.conda
- hash:
- md5: 52a472fd24ee3b3d398fc42760398c6f
- sha256: f692d88ef78a92ac8009d48b78e9be008b991803f0cc11bb93ed2220ebbfaa62
- category: main
- optional: false
-- name: rapids-logger
- version: 0.2.3
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=13'
- libstdcxx: '>=13'
- url: https://conda.anaconda.org/rapidsai/linux-64/rapids-logger-0.2.3-h98325ef_0.conda
- hash:
- md5: 81257f29bfcc1e58f0405d7bc9feb309
- sha256: 6d0416d415128ce1219fe9d094447b12a612e02e7317006b3dcb3050a0f7f8d9
- category: main
- optional: false
- name: rdma-core
version: '65.0'
manager: conda
@@ -6176,10 +4673,10 @@ package:
platform: linux-64
dependencies:
libre2-11: 2025.11.05
- url: https://conda.anaconda.org/conda-forge/linux-64/re2-2025.11.05-hbf548d4_3.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/re2-2025.11.05-hbf548d4_4.conda
hash:
- md5: eac6d593030ba95beda012452003ed53
- sha256: a77bfa9d62f1df59d46e8de8961ebc6b618ea055ccd9e3f050ace4d10f459ff8
+ md5: f05fcc2734b47231b9a5623c6d4f43ac
+ sha256: 54a40d376a64b91b1a91decb1d4986f189cfb26fd4cdaddd850e5aafa092500d
category: main
optional: false
- name: readline
@@ -6227,28 +4724,6 @@ package:
sha256: 3d6ba2c0fcdac3196ba2f0615b4104e532525ffa1335b50a2878be5ff488814a
category: main
optional: false
-- name: rmm
- version: 26.04.00
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.28,<3.0.a0'
- _python_abi3_support: 1.*
- cpython: '>=3.11'
- cuda-cudart: ''
- cuda-python: '>=12.9.2,<13.0'
- cuda-version: '>=12,<13.0a0'
- libgcc: '>=14'
- librmm: '>=26.4.0,<26.5.0a0'
- libstdcxx: '>=14'
- numpy: '>=1.23,<3.0'
- python: ''
- url: https://conda.anaconda.org/rapidsai/linux-64/rmm-26.04.00-cuda12_cp311_abi3_260408_48b36cc6.conda
- hash:
- md5: 347f02e4a94831af80076fd6157776ac
- sha256: 690cdf5168023a87228c083da2ddc0a83302d9783904cff946a0f23541020f78
- category: main
- optional: false
- name: s2n
version: 1.7.7
manager: conda
@@ -6364,24 +4839,6 @@ package:
sha256: 90861e7f903ad48a25a8f5d0d4a74b010b517c47026ca32fe385e833542a1055
category: main
optional: false
-- name: scikit-misc
- version: 0.5.3
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- libgcc: '>=15'
- libgfortran: ''
- libgfortran5: '>=15.3.0'
- numpy: '>=1.25,<3'
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/scikit-misc-0.5.3-py312hc3447ac_0.conda
- hash:
- md5: 5a3d17773efbf1286b9a4732c9bae445
- sha256: 3d554ffb3eba5c38f2e89ac325dc03362c1b6f24ecd509d4869c4758f2012916
- category: main
- optional: false
- name: scipy
version: 1.16.3
manager: conda
@@ -6572,29 +5029,17 @@ package:
category: main
optional: false
- name: snappy
- version: 1.2.2
+ version: 1.3.1
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
libgcc: '>=15'
libstdcxx: '>=15'
- url: https://conda.anaconda.org/conda-forge/linux-64/snappy-1.2.2-h34e00bb_2.conda
+ url: https://conda.anaconda.org/conda-forge/linux-64/snappy-1.3.1-h34e00bb_0.conda
hash:
- md5: 2b7e31ffbd476eecf33b3b833692339c
- sha256: 8635d10c7f127808b09c0c8e77b21a255829e0934e49326548c949f7e0551f09
- category: main
- optional: false
-- name: sortedcontainers
- version: 2.4.0
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.9'
- url: https://conda.anaconda.org/conda-forge/noarch/sortedcontainers-2.4.0-pyhd8ed1ab_1.conda
- hash:
- md5: 0401a17ae845fa72c7210e206ec5647d
- sha256: d1e3e06b5cf26093047e63c8cc77b70d970411c5cbc0cb1fad461a8a8df599f7
+ md5: 9fe09f49cf48a5e73ef0c8dbfad570ba
+ sha256: a724d5312f1c8fe84be83b176b20fa171b3106cdfec1648cc3f32888397ecb09
category: main
optional: false
- name: stack_data
@@ -6687,18 +5132,6 @@ package:
sha256: ea35780ec77e2431bcfc193e75ddcc3ec1ae79a34464469b0b2186d93035c9e7
category: main
optional: false
-- name: tblib
- version: 3.2.2
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.10'
- url: https://conda.anaconda.org/conda-forge/noarch/tblib-3.2.2-pyhcf101f3_0.conda
- hash:
- md5: f88bb644823094f436792f80fba3207e
- sha256: 6b549360f687ee4d11bf85a6d6a276a30f9333df1857adb0fe785f0f8e9bcd60
- category: main
- optional: false
- name: texttable
version: 1.7.2
manager: conda
@@ -6712,28 +5145,28 @@ package:
category: main
optional: false
- name: threadpoolctl
- version: 3.7.0
+ version: 3.8.0
manager: conda
platform: linux-64
dependencies:
python: '>=3.11'
- url: https://conda.anaconda.org/conda-forge/noarch/threadpoolctl-3.7.0-pyhc455866_0.conda
+ url: https://conda.anaconda.org/conda-forge/noarch/threadpoolctl-3.8.0-pyhc455866_0.conda
hash:
- md5: 8c05910e780502e362d658f00f5ec10f
- sha256: a32d17fe979e59fe1295875b5d4f0431a76dfbbfcee5b01e657a8aa3c4a1b5d1
+ md5: a84c9b62a40f5d6f1c07df8117ae770b
+ sha256: 6dde58ec2e194419a7381622a3177fa9d4c424823e9c65b04b905892ce7d7779
category: main
optional: false
- name: tifffile
- version: 2026.10.15
+ version: 2026.11.8
manager: conda
platform: linux-64
dependencies:
numpy: '>=2.1'
python: '>=3.12'
- url: https://conda.anaconda.org/conda-forge/noarch/tifffile-2026.10.15-pyhc364b38_0.conda
+ url: https://conda.anaconda.org/conda-forge/noarch/tifffile-2026.11.8-pyhc364b38_0.conda
hash:
- md5: 69311a22cf97c3c9840108555b83e4f7
- sha256: 74205e597f33d5aa55dc535481cff7d7341b8026c1698be0d0c62f1e7c42110f
+ md5: da6fc49f7ac67a900f93a51ddf15d585
+ sha256: d5505801ed156186a416ac650406c55951e1fc7bb8a31c5c7442d65c415a3e99
category: main
optional: false
- name: tk
@@ -6750,18 +5183,6 @@ package:
sha256: a1a241d172c1ccab067ba245206dd048bc3c2d1b84504b53c9468e99adfc16a1
category: main
optional: false
-- name: toolz
- version: 1.2.0
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.11'
- url: https://conda.anaconda.org/conda-forge/noarch/toolz-1.2.0-pyh5ded981_0.conda
- hash:
- md5: 17f8db9088267a3cb38d56cc22d268ae
- sha256: 582d3ab0298347c5f46439c7ce834c7152eb43b0cce1e229568de7235cddad13
- category: main
- optional: false
- name: tornado
version: '6.6'
manager: conda
@@ -6802,40 +5223,21 @@ package:
sha256: 03dba5917f944c6684ab44c81daacac1624cd148e4b2cae215dcec594a210c48
category: main
optional: false
-- name: treelite
- version: 4.8.1
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- _openmp_mutex: '>=4.5'
- libgcc: '>=15'
- libstdcxx: '>=15'
- numpy: ''
- packaging: ''
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- scipy: ''
- url: https://conda.anaconda.org/conda-forge/linux-64/treelite-4.8.1-py312h253b9dd_0.conda
- hash:
- md5: 9c467f11b09cc4f5117186283406efb9
- sha256: a4c122f2f622bc0b1c0c306bceb789c520813e80323f3d07e77fd36e2fe90523
- category: main
- optional: false
- name: typer
- version: 0.27.3
+ version: 0.29.0
manager: conda
platform: linux-64
dependencies:
annotated-doc: '>=0.0.2'
colorama: ''
+ pydantic: '>=2.5.3'
python: '>=3.11'
rich: '>=13.8.0'
shellingham: '>=1.3.0'
- url: https://conda.anaconda.org/conda-forge/noarch/typer-0.27.3-pyh5ded981_0.conda
+ url: https://conda.anaconda.org/conda-forge/noarch/typer-0.29.0-pyh5ded981_0.conda
hash:
- md5: 3251bd81ef66ec6432b3159944c70d57
- sha256: 50cef25b46782fbb8291427459ed60c701165b0b70bdb3aebae2503e9915f66a
+ md5: 50b1de5cb639dff936c40b1577015f9a
+ sha256: e0e98c372dc47488d8c49c47800c8ce8da29eb6a66ebd4befd03c7b6fb71f44d
category: main
optional: false
- name: typing-extensions
@@ -6887,43 +5289,19 @@ package:
category: main
optional: false
- name: ucx
- version: 1.20.1
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- _openmp_mutex: '>=4.5'
- libgcc: '>=14'
- libstdcxx: '>=14'
- rdma-core: '>=61.0'
- url: https://conda.anaconda.org/conda-forge/linux-64/ucx-1.20.1-hbe80e26_0.conda
- hash:
- md5: 7d06bc10996e75c90b8cd7631b5dcf6c
- sha256: 8f0ac4a92e4674eb4c87cdfc59d2fc7c2d0d1696689202eeb62ef715d82ce711
- category: main
- optional: false
-- name: ucxx
- version: 0.49.00
+ version: 1.22.0
manager: conda
platform: linux-64
dependencies:
__glibc: '>=2.28,<3.0.a0'
- _python_abi3_support: 1.*
- cpython: '>=3.11'
- cuda-core: '>=0.3.2'
- cuda-version: '>=12,<13.0a0'
- libgcc: '>=14'
- libstdcxx: '>=14'
- libucxx: 0.49.0.*
- numpy: '>=1.23,<3.0'
- nvidia-ml-py: '>=12'
- python: ''
- rmm: '>=26.4.0,<26.5.0a0'
- ucx: '>=1.18.0,<1.21.0'
- url: https://conda.anaconda.org/rapidsai/linux-64/ucxx-0.49.00-cuda12_cp311_abi3_260408_8d47a9ff.conda
+ _openmp_mutex: '>=4.5'
+ libgcc: '>=15'
+ libstdcxx: '>=15'
+ rdma-core: '>=65.0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/ucx-1.22.0-h6cc3e34_1.conda
hash:
- md5: 4e44176d2cfc7f9941dc3089df22c9dc
- sha256: 4145a241563f1b9bafe55e674f44fd623c33ffef7bacb1581ae895297c237605
+ md5: 6645226631e892dab56fbf03eb635049
+ sha256: e8fb16ed2033c66bc81fb36003e8bc0a21db9232aa5fec8f3a7ae5bd97e5085c
category: main
optional: false
- name: umap-learn
@@ -7385,13 +5763,13 @@ package:
platform: linux-64
dependencies:
__glibc: '>=2.17,<3.0.a0'
- libgcc: '>=14'
- xorg-libx11: '>=1.8.12,<2.0a0'
- xorg-libxext: '>=1.3.6,<2.0a0'
- url: https://conda.anaconda.org/conda-forge/linux-64/xorg-libxxf86vm-1.1.7-hb03c661_0.conda
+ libgcc: '>=15'
+ xorg-libx11: '>=1.8.14,<2.0a0'
+ xorg-libxext: '>=1.3.7,<2.0a0'
+ url: https://conda.anaconda.org/conda-forge/linux-64/xorg-libxxf86vm-1.1.7-h7cc23a3_1.conda
hash:
- md5: 665d152b9c6e78da404086088077c844
- sha256: 64db17baaf36fa03ed8fae105e2e671a7383e22df4077486646f7dbf12842c9f
+ md5: ac05771f924d124349ef9cc3140e9876
+ sha256: 264ddaeba8de4fe17fb25d10c6f6c2aac5d8d41d8883a0976dae59b528ef9520
category: main
optional: false
- name: xorg-xorgproto
@@ -7407,18 +5785,6 @@ package:
sha256: 8e9c7ca4fcfe9f0c300e4aa853b79c25411fc4e470decfe8f909b74f60dfb6dd
category: main
optional: false
-- name: xyzservices
- version: 2026.9.1
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.9'
- url: https://conda.anaconda.org/conda-forge/noarch/xyzservices-2026.9.1-pyhd8ed1ab_0.conda
- hash:
- md5: 5d7c149ba9c728d4951f3fe6d3ce0f84
- sha256: b9439f1716ff580ce56e9b4a4df7ea92e5ed2bdc9567d25cdff223d62c3183a8
- category: main
- optional: false
- name: yaml
version: 0.2.5
manager: conda
@@ -7432,24 +5798,6 @@ package:
sha256: d164dfa75ecd538f6fd68765defcc06aa875bc697b9b215362d79a2a73125dd0
category: main
optional: false
-- name: yarl
- version: 1.25.1
- manager: conda
- platform: linux-64
- dependencies:
- __glibc: '>=2.17,<3.0.a0'
- idna: '>=2.0'
- libgcc: '>=15'
- multidict: '>=4.0'
- propcache: '>=0.2.1'
- python: '>=3.12,<3.13.0a0'
- python_abi: 3.12.*
- url: https://conda.anaconda.org/conda-forge/linux-64/yarl-1.25.1-py312hfd18d64_0.conda
- hash:
- md5: 91bb35918653d559dbd514c9064e512b
- sha256: 56cee1531e3d0a0a61bfc78a87ea294b7b9cb720221a99f274732fd863b983e6
- category: main
- optional: false
- name: zarr
version: 3.4.2
manager: conda
@@ -7469,18 +5817,6 @@ package:
sha256: 69fe66c367849ec39ebf9c47aa329cb20c3488cbff572b287d2d609fc1bcdf61
category: main
optional: false
-- name: zict
- version: 3.0.0
- manager: conda
- platform: linux-64
- dependencies:
- python: '>=3.9'
- url: https://conda.anaconda.org/conda-forge/noarch/zict-3.0.0-pyhd8ed1ab_1.conda
- hash:
- md5: e52c2ef711ccf31bb7f70ca87d144b9e
- sha256: 5488542dceeb9f2874e726646548ecc5608060934d6f9ceaa7c6a48c61f9cc8d
- category: main
- optional: false
- name: zipp
version: 4.1.2
manager: conda
@@ -7555,13 +5891,14 @@ package:
category: main
optional: false
- name: cell2cell
- version: 0.8.4
+ version: 0.9.0
manager: pip
platform: linux-64
dependencies:
gseapy: '>=1.0.3'
kneed: '*'
matplotlib: '>=3.2.0'
+ natsort: '*'
networkx: '>=2.3'
numpy: '>=1.16'
openpyxl: '>=2.6.2'
@@ -7575,13 +5912,13 @@ package:
tqdm: '*'
umap-learn: '*'
xlrd: '>=1.1'
- url: https://files.pythonhosted.org/packages/39/40/d0c5b144ff8f880e4e03da35969ce6ff055cde3e6544b722eda24f19c357/cell2cell-0.8.4-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/6f/70/7b36729239f848ea264c2493c52e7247d8eaf8f6e0582767125a395ea40a/cell2cell-0.9.0-py3-none-any.whl
hash:
- sha256: 506dcffadafc06d2dc3358213e8df5ca16ab5cca1c8f48c03640a76f341d56ca
+ sha256: 4a2d2e3cfb7706def00183fe8205ae08a1106bf166b33ea6c23d9b842098f4d6
category: main
optional: false
- name: cellrank
- version: 2.3.2
+ version: 2.3.3
manager: pip
platform: linux-64
dependencies:
@@ -7602,9 +5939,9 @@ package:
seaborn: '>=0.13'
session-info2: '>=0.2'
wrapt: '>=1.16'
- url: https://files.pythonhosted.org/packages/63/02/3f07b87bdcc0354f7b4bef973b2ecc60676035ec04754c4cc7457ff9a4c1/cellrank-2.3.2-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/71/9d/45fa014e466d80ca492f4a839b3b416343c6ddd9e08fb03e859fc3e04697/cellrank-2.3.3-py3-none-any.whl
hash:
- sha256: 7a29e03cabac4a7bc80335e48341a10f69383d36283ff84dc8e0e9a086932cf5
+ sha256: 351578a18beb36bbba24e1b79cb0545800ea34e0c960497d7878ae8f297a1f8b
category: main
optional: false
- name: celltypist
@@ -7625,6 +5962,16 @@ package:
sha256: 3953f96a4476ed1dd2807446d97d3945e8905bbdac6bfef0d20b84a9a00d1101
category: main
optional: false
+- name: click
+ version: 8.5.1
+ manager: pip
+ platform: linux-64
+ dependencies: {}
+ url: https://files.pythonhosted.org/packages/31/7d/164b7ffc8624fe9efb4221bf9af485d5c1e41b9ae118045d2b6dc5457d8b/click-8.5.1-py3-none-any.whl
+ hash:
+ sha256: 8bb5376d2006950386f8c6427613334a2acadc5f559cb80e638d00b4df957e37
+ category: main
+ optional: false
- name: deprecated
version: 3.0.1
manager: pip
@@ -7637,13 +5984,13 @@ package:
category: main
optional: false
- name: filelock
- version: 4.2.9
+ version: 4.2.10
manager: pip
platform: linux-64
dependencies: {}
- url: https://files.pythonhosted.org/packages/00/05/76184fe08339723d643ab47887a8e4de8f2921ca4e008617ab7490af7857/filelock-4.2.9-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/ef/90/5d0096a325aa0f83ba7a3968bfb485b67e32c947bdf9cd0fcb4b485be718/filelock-4.2.10-py3-none-any.whl
hash:
- sha256: a379c50a729e5b68c8519a4d26489f755d9769960a236ecd838190e2a78db2e9
+ sha256: 9479aaeb4db20b5a53e3da28493edd24720b66b8a0b84030aa56ccdf243d6857
category: main
optional: false
- name: future
@@ -7754,22 +6101,23 @@ package:
category: main
optional: false
- name: liana
- version: 1.7.1
+ version: 2.1.1
manager: pip
platform: linux-64
dependencies:
- anndata: '>=0.7.4'
- docrep: '>=0.3.1'
- mudata: '*'
+ anndata: '>=0.13'
+ fast-array-utils: '*'
+ mudata: '>=0.4'
numba: '*'
pandas: '*'
plotnine: '>=0.10.1'
- scanpy: '*'
+ scanpy: '>=1.12'
+ scverse-misc: '>=0.1.4'
session-info2: '*'
- tqdm: '>=4.0'
- url: https://files.pythonhosted.org/packages/11/35/24bcb5134ad91fb9787c85c0c1ef5f81b1b87adf9dfb29e8edb5e0050ca2/liana-1.7.1-py3-none-any.whl
+ tqdm: '>=4'
+ url: https://files.pythonhosted.org/packages/97/8c/62c88ea51e543cd386cad54de71070322d4a20ed6c899a006cd50799ebf1/liana-2.1.1-py3-none-any.whl
hash:
- sha256: f4eead6df3f7fd07648fda77a50b50673270e1fca19030a7af802d57b84eb2e6
+ sha256: 195253f4fd3f9317e6e41aee114bd29d7a3081cdce5a847a947db1c763472864
category: main
optional: false
- name: lightning
@@ -7898,14 +6246,14 @@ package:
category: main
optional: false
- name: numpy-groupies
- version: 0.16.3
+ version: 0.16.5
manager: pip
platform: linux-64
dependencies:
numpy: '*'
- url: https://files.pythonhosted.org/packages/d6/e2/08809106723ddb004e7af0399b36848667a9f58b40bb85008f5dfc5752d1/numpy_groupies-0.16.3-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/30/75/c7638e6bc615c5bbd044c2e055bb112545767873753546a84d5fb60136e8/numpy_groupies-0.16.5-py3-none-any.whl
hash:
- sha256: 6dd37aff8dc1a172d00589ea93f05d398df9f53f7370266208875ca21542cafb
+ sha256: a9e6cc17d2b93f40ce0d07348091623ee1f5b8c531177721301cb24b602afbfd
category: main
optional: false
- name: nvidia-cublas-cu12
@@ -8262,27 +6610,15 @@ package:
sha256: 71f95c7b22f25c4f91cc659e1734bcdc6c69c8b66543f2ed2d78dbe29ce2f167
category: main
optional: false
-- name: rapids-singlecell-cu12
- version: 0.15.2
+- name: scikit-misc
+ version: 0.5.3
manager: pip
platform: linux-64
dependencies:
- anndata: '>=0.10.0'
- docrep: '*'
- matplotlib: '>=3.4'
- natsort: '*'
- numpy: '>=1.17.0'
- pandas: '*'
- scanpy: '>=1.10.0'
- scikit-image: '*'
- scikit-learn: '*'
- scikit-misc: '>=0.1.3'
- scipy: '>=1.4'
- statsmodels: '>=0.12.0'
- tqdm: '*'
- url: https://files.pythonhosted.org/packages/4c/d8/c7c845fa8444285068af5cdf9477d59d8644577c650ab4c1c9f063ecff5e/rapids_singlecell_cu12-0.15.2-cp312-abi3-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl
+ numpy: '>=1.26.4'
+ url: https://files.pythonhosted.org/packages/ec/52/3ba0bb75fe3d9a6d2edbe82463dfbc05f5f665ffa28fb4f123083a673071/scikit_misc-0.5.3-cp312-cp312-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl
hash:
- sha256: 29cf5ef827220dbd99ef35307bc257db5a3587e4de81db6bc7ddc7bb7423f5ba
+ sha256: 1861b2fa1619dd6645154cdb7d9696d71f0a0e7ef8f5201f1e7bacf21507c858
category: main
optional: false
- name: scvelo
@@ -8306,32 +6642,33 @@ package:
category: main
optional: false
- name: scvi-tools
- version: 1.4.3
+ version: 1.5.2
manager: pip
platform: linux-64
dependencies:
anndata: '*'
docrep: '>=0.3.2'
- lightning: '>=2.0'
+ joblib: '*'
+ lightning: '>=2'
ml-collections: '*'
mudata: '*'
- numba: '>=0.60.0'
+ numba: '>=0.60'
numpy: '*'
pandas: '*'
pyro-ppl: '*'
rich: '*'
- scanpy: '>=1.10'
+ scanpy: '>=1.11'
scikit-learn: '*'
scipy: '*'
- sparse: '>=0.14.0'
+ sparse: '>=0.14'
tensorboard: '*'
torch: '*'
torchmetrics: '*'
tqdm: '*'
xarray: '*'
- url: https://files.pythonhosted.org/packages/ed/87/ba97e56c8aa0e07ac68a458df5dae9c696cd1a5ffe2b80c9259bb8f7a5d6/scvi_tools-1.4.3-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/89/2c/b672e659ee21e6691c12d6206e1f5c1a6e49ee96c549e77adaf312178b34/scvi_tools-1.5.2-py3-none-any.whl
hash:
- sha256: 635a21db462d27731bc3b1e569203cdfe491ca656546da2fbf139b5574a54960
+ sha256: ecc33564f325b3e6053976725f7eb242060edbbfbc2d9cf13b100405eca790c9
category: main
optional: false
- name: sparse
@@ -8414,15 +6751,15 @@ package:
category: main
optional: false
- name: tensorly
- version: 0.9.0
+ version: 0.10.0
manager: pip
platform: linux-64
dependencies:
numpy: '*'
scipy: '*'
- url: https://files.pythonhosted.org/packages/c3/ad/83380fb3c85ddf68c509059cf328374c1878f87382070475abc72867c7ba/tensorly-0.9.0-py3-none-any.whl
+ url: https://files.pythonhosted.org/packages/17/02/0269a1573b9f2ea6365e12b87debaf8b2cdad8e3fc4f84d9f4dc6be68adb/tensorly-0.10.0-py3-none-any.whl
hash:
- sha256: a393022ff1616a36c021496f46a955d94234d6bc8c42793f477a6be8849f4873
+ sha256: 90becb3d28023638ec60d19fbf8d87735b4ce1537fe3eb881897d69e107b882d
category: main
optional: false
- name: torch
diff --git a/envs/cellquorum-gpu.yml b/envs/cellquorum-gpu.yml
index 1a6d49ef..69d497c8 100644
--- a/envs/cellquorum-gpu.yml
+++ b/envs/cellquorum-gpu.yml
@@ -1,45 +1,44 @@
# Linux x86_64 / Python 3.12 / CUDA 12.9 workstation profile.
-# Validated scope and unresolved metadata/locking limits: envs/README.md.
+# RAPIDS uses the separate CUDA worker in rapids-worker-requirements.txt (envs/README.md).
# Install CellQuorum separately from its built wheel after creating this environment.
name: cellquorum-gpu
channels:
-- rapidsai
- conda-forge
-- nvidia
- bioconda
dependencies:
- python=3.12
- pip
- git
- cython
-- numpy>=1.24
+- numpy=2.3.5
- scipy=1.16.3
-- pandas>=2.1
+- pandas=3.1.1
- scikit-learn>=1.4
- statsmodels>=0.14
- numba>=0.59
-- anndata>=0.12,<0.13
+- anndata=0.13.4
# mudata (a scvi-tools and liana dependency) imports Settings from scverse-misc, which needs scverse-misc's
# `settings` extra: pydantic-settings and python-dotenv. A conda install of scverse-misc does not bring
# extras, so without these two `import mudata`, `import scvi` and `import liana` fail.
- pydantic-settings
+# LIANA imports pooch when loading its orthology resources.
+- pooch>=1.8,<2
- python-dotenv
-- scanpy>=1.10
+- scanpy=1.12.5
- h5py>=3.10
- zarr>=2.17
- natsort>=8.4
-- umap-learn>=0.5
+- umap-learn=0.5.12
- pynndescent>=0.5
- python-igraph>=0.11
- leidenalg>=0.10
- louvain>=0.8
-- scikit-misc>=0.3
- petsc4py>=3.22
- slepc4py>=3.22
- harmonypy>=0.0.10
- bbknn>=1.6
- scrublet>=0.2
-- decoupler-py>=1.8
+- decoupler-py=2.2.0
- networkx>=3.2
- joblib>=1.3
- tqdm>=4.66
@@ -56,25 +55,18 @@ dependencies:
- pydantic>=2.7
- typer>=0.12
- rich>=13.0
-- cuda-version=12.9
-- cupy=14.1.1
-- cudf=26.4
-- cuml=26.4
-- cugraph=26.4
-- cuvs=26.4
- pip:
- torch @ https://download-r2.pytorch.org/whl/cu129/torch-2.8.0%2Bcu129-cp312-cp312-manylinux_2_28_x86_64.whl#sha256=02c7258e917f3043c978b53acf6f02b818db0d0d85db0e58ae578af333b9b4e2
- pygpcca @ git+https://github.com/msmdev/pygpcca.git@773c463cc5043a5d40e6f31bbbc20f4e49587008
- - scvi-tools==1.4.3
+ - scikit-misc==0.5.3
+ - scvi-tools==1.5.2
- phate==2.0.0
- celltypist==1.7.1
- pycirclize==1.10.1
- plotly==6.7.0
- - rapids-singlecell-cu12==0.15.2
- - liana==1.7.1
- - cell2cell==0.8.4
- - tensorly==0.9.0
+ - liana==2.1.1
+ - cell2cell==0.9.0
+ - tensorly==0.10.0
- scvelo==0.3.4
- - cellrank==2.3.2
+ - cellrank==2.3.3
- palantir==1.4.4
- - anndata==0.12.19
diff --git a/envs/ci-constraints.txt b/envs/ci-constraints.txt
index 6c64357d..f9d13d4c 100644
--- a/envs/ci-constraints.txt
+++ b/envs/ci-constraints.txt
@@ -1,11 +1,8 @@
-# The library versions the CI typecheck and test jobs are clean against. Unpinned,
-# `pip install -e .[dev]` takes whatever is newest (anndata 0.13 retyped X and obs, pandas 3 reads
-# booleans and strings differently, numpy 2.5 changed nditer, matplotlib 3.11 dropped ignores), and
-# the gate and the tests report failures that are not in this code. Move these together, in their
-# own commit, when both are rechecked against newer releases.
-anndata==0.12.10
-pandas==2.3.3
+# Versions exercised together by the CI test and typecheck jobs.
+anndata==0.13.4
+pandas==3.1.1
numpy==2.4.6
scipy==1.16.3
matplotlib==3.10.8
scanpy==1.12.5
+umap-learn==0.5.12
diff --git a/envs/cytotrace2-requirements.txt b/envs/cytotrace2-requirements.txt
new file mode 100644
index 00000000..aa35c66d
--- /dev/null
+++ b/envs/cytotrace2-requirements.txt
@@ -0,0 +1,12 @@
+# Independent CytoTRACE 2 CPU worker for the extended lineage reference.
+cytotrace2-py==1.1.0.4
+numpy==1.26.4
+pandas==2.3.3
+scanpy==1.11.5
+anndata==0.12.6
+# CytoTRACE 2 uses pkg_resources to locate the model ensemble.
+setuptools==80.9.0
+scipy==1.17.1
+scikit-learn==1.9.1
+# Official CPU wheel for the Linux / WSL Python 3.12 reference worker.
+torch @ https://download-r2.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-manylinux_2_28_x86_64.whl#sha256=cb9a8ba8137ab24e36bf1742cb79a1294bd374db570f09fc15a5e1318160db4e ; sys_platform == "linux" and platform_machine == "x86_64" and python_version == "3.12"
diff --git a/envs/lineage-constraints.txt b/envs/lineage-constraints.txt
index d740040f..495bd0b3 100644
--- a/envs/lineage-constraints.txt
+++ b/envs/lineage-constraints.txt
@@ -1,6 +1,15 @@
-# Versions used for the mouse pancreatic lineage reference.
+# Constraints for the mouse pancreatic lineage reference.
numpy==2.3.5
scipy==1.16.3
scanpy==1.12.5
scvelo==0.3.4
-cellrank==2.3.2
+# 2.3.3 supports the HDF5 registration changes in AnnData 0.13.4.
+cellrank==2.3.3
+pandas==3.1.1
+anndata==0.13.4
+# PyGPCCA 1.0.4 on PyPI requires Jinja2 3.0.3. The pinned upstream source
+# used by our CPU and GPU environments accepts the report renderer's Jinja2.
+pygpcca @ git+https://github.com/msmdev/pygpcca.git@773c463cc5043a5d40e6f31bbbc20f4e49587008
+# Match the manuscript renderer and the diagnostic PNGs of the lineage baseline.
+matplotlib==3.10.9
+pillow==12.3.0
diff --git a/envs/rapids-worker-requirements.txt b/envs/rapids-worker-requirements.txt
new file mode 100644
index 00000000..2be80b42
--- /dev/null
+++ b/envs/rapids-worker-requirements.txt
@@ -0,0 +1,12 @@
+# Linux x86_64, Python 3.12, CUDA 12. Install separately from CellQuorum.
+rapids-singlecell-cu12[rapids]==0.18.1
+cupy-cuda12x==14.2.0
+cudf-cu12==26.10.2
+cugraph-cu12==26.10.0
+cuml-cu12==26.10.0
+cuvs-cu12==26.10.0
+pandas==3.0.6
+anndata==0.13.4
+numpy==2.3.5
+scipy==1.16.3
+scanpy==1.12.5
diff --git a/envs/scc-signaling-constraints.txt b/envs/scc-signaling-constraints.txt
index 99531a0c..08cc44b4 100644
--- a/envs/scc-signaling-constraints.txt
+++ b/envs/scc-signaling-constraints.txt
@@ -1,5 +1,5 @@
-# Direct Python dependencies from the paired SCC installed-wheel run, Python 3.12.
-liana==1.10.0
+# Python 3.12 inputs for the paired SCC signaling recipe.
+liana==2.1.1
cell2cell==0.9.0
tensorly==0.10.0
decoupler==2.2.0
diff --git a/pyproject.toml b/pyproject.toml
index 40b1f0c6..4f9c2469 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -44,6 +44,8 @@ dependencies = [
"pydantic>=2.7",
"PyYAML>=6.0",
"numpy>=1.24",
+ # 0.6.0 fails Numba compilation when Scanpy requests approximate CPU neighbors.
+ "umap-learn!=0.6.0",
"pandas>=2.1",
"scipy>=1.11",
# Earlier versions call getnnz on the sparse arrays returned by PyNNDescent 0.7.
@@ -65,7 +67,7 @@ dependencies = [
]
[project.optional-dependencies]
-signaling = ["liana>=1.10,<2", "cell2cell>=0.9,<1", "decoupler>=2.2,<3"]
+signaling = ["liana>=1.10,<3", "pooch>=1.8,<2", "cell2cell>=0.9,<1", "decoupler>=2.2,<3"]
enrichment = [
"decoupler>=2.1.6,<3",
@@ -156,7 +158,7 @@ where = ["src"]
[tool.setuptools.data-files]
"share/cellquorum/profiles" = ["configs/profiles/*.yaml"]
-"share/cellquorum/benchmarks" = ["docs/benchmarks/study-coverage.csv", "docs/benchmarks/reference-results.json", "docs/benchmarks/scientific-validation.csv", "docs/benchmarks/scientific-validation.md", "docs/benchmarks/scientific-validation.svg", "docs/benchmarks/reference-studies.md", "docs/benchmarks/pbmc-identity.md", "envs/scc-signaling-constraints.txt", "envs/lineage-constraints.txt"]
+"share/cellquorum/benchmarks" = ["docs/benchmarks/study-coverage.csv", "docs/benchmarks/reference-results.json", "docs/benchmarks/scientific-validation.csv", "docs/benchmarks/scientific-validation.md", "docs/benchmarks/scientific-validation.svg", "docs/benchmarks/reference-studies.md", "docs/benchmarks/pbmc-identity.md", "envs/scc-signaling-constraints.txt", "envs/lineage-constraints.txt", "envs/cytotrace2-requirements.txt"]
"share/cellquorum/benchmarks/pbmc-identity" = ["docs/benchmarks/pbmc-identity/*"]
"share/cellquorum/benchmarks/scc-programs" = ["docs/benchmarks/scc-programs/*"]
"share/cellquorum/workflows/helpers" = ["scripts/prepare_scc.R", "scripts/prepare_pancreas_atlas.py", "scripts/prepare_perturbseq.py", "scripts/verify_reference_study.py", "scripts/study_coverage.py", "scripts/validate_study_results.py", "scripts/fetch_scc_resources.py", "scripts/prepare_kang_specimens.py", "scripts/prepare_hao_blind.py", "scripts/evaluate_pbmc_identity.py", "scripts/figure_scc_programs.py"]
diff --git a/scripts/check_environment_imports.py b/scripts/check_environment_imports.py
index dc2ca4af..1e2f96c7 100644
--- a/scripts/check_environment_imports.py
+++ b/scripts/check_environment_imports.py
@@ -21,7 +21,10 @@
from __future__ import annotations
import importlib
+import os
+import subprocess
import sys
+from pathlib import Path
CORE = [
"anndata",
@@ -60,16 +63,11 @@
"cellquorum",
]
GPU = ["cupy", "cuml", "cugraph", "cuvs", "rapids_singlecell"]
-PROFILES = {"core": CORE, "gpu": CORE + GPU}
-def main() -> int:
- profile = sys.argv[1] if len(sys.argv) > 1 else "core"
- if profile not in PROFILES:
- print(f"unknown profile {profile!r}; use one of {sorted(PROFILES)}")
- return 2
+def check_imports(names: list[str]) -> bool:
failed = []
- for name in PROFILES[profile]:
+ for name in names:
try:
module = importlib.import_module(name)
except Exception as error: # noqa: BLE001 - any failure to import is the finding
@@ -77,17 +75,43 @@ def main() -> int:
print(f"FAIL {name}: {type(error).__name__}: {str(error)[:100]}")
else:
print(f"ok {name} {getattr(module, '__version__', '')}")
- if profile == "gpu":
- import cupy
- import torch
-
- devices = cupy.cuda.runtime.getDeviceCount()
- available = torch.cuda.is_available()
- print(f"GPU devices visible to cupy: {devices}; torch.cuda.is_available(): {available}")
- if devices < 1 or not available:
- failed.append("gpu-device")
- print(f"\n{len(PROFILES[profile]) - len(failed)} of {len(PROFILES[profile])} imports ok")
- return 1 if failed else 0
+ print(f"{len(names) - len(failed)} of {len(names)} imports ok", flush=True)
+ return not failed
+
+
+def check_rapids() -> bool:
+ if not check_imports(GPU):
+ return False
+ import cupy
+
+ devices = cupy.cuda.runtime.getDeviceCount()
+ print(f"GPU devices visible to CuPy: {devices}")
+ return devices > 0
+
+
+def main() -> int:
+ profile = sys.argv[1] if len(sys.argv) > 1 else "core"
+ if profile == "rapids-worker":
+ return int(not check_rapids())
+ if profile not in {"core", "gpu"}:
+ print(f"unknown profile {profile!r}; use core or gpu")
+ return 2
+ core_ok = check_imports(CORE)
+ if profile == "core":
+ return int(not core_ok)
+ worker = os.environ.get("CELLQUORUM_RAPIDS_PYTHON")
+ if worker:
+ result = subprocess.run([worker, "-I", str(Path(__file__).resolve()), "rapids-worker"])
+ rapids_ok = result.returncode == 0
+ else:
+ rapids_ok = check_rapids()
+ if not core_ok:
+ return 1
+ import torch
+
+ available = torch.cuda.is_available()
+ print(f"torch.cuda.is_available(): {available}")
+ return int(not (rapids_ok and available))
if __name__ == "__main__":
diff --git a/scripts/compare_scc_paired_ranks.py b/scripts/compare_scc_paired_ranks.py
new file mode 100644
index 00000000..b4bb1370
--- /dev/null
+++ b/scripts/compare_scc_paired_ranks.py
@@ -0,0 +1,143 @@
+"""Recompute paired SCC rank evidence from the matched LIANA library exports."""
+
+from pathlib import Path
+
+import numpy as np
+import pandas as pd
+from scipy.stats import binomtest, false_discovery_control
+
+ROOT = Path(__file__).resolve().parents[1] / "docs/benchmarks/pandas-anndata-review"
+KEY = ["source", "target", "ligand_complex", "receptor_complex"]
+
+
+def main() -> None:
+ ranks = pd.read_csv(ROOT / "scc-magnitude-interactions.csv.gz")
+ libraries = (
+ pd.read_csv(ROOT / "scc-panel-measurements.csv")[["sample", "donor", "group"]]
+ .dropna()
+ .drop_duplicates()
+ )
+ ranks = ranks.merge(libraries, on="sample", how="left", validate="many_to_one")
+ if ranks["donor"].isna().any():
+ raise ValueError("Missing donor metadata for scored interaction")
+ values = ["base_percentile", "next_percentile"]
+ paired = ranks.pivot(index=[*KEY, "donor"], columns="group", values=values).dropna()
+ differences = paired.xs("Tumor", axis=1, level="group") - paired.xs(
+ "Normal", axis=1, level="group"
+ )
+ groups = differences.groupby(KEY)
+ summary = pd.DataFrame({"pairs": groups.size()})
+ for name, column in zip(("baseline", "candidate"), values, strict=True):
+ summary[name + "_median_delta_pp"] = groups[column].median()
+ stronger = differences[column].lt(0).groupby(KEY).sum()
+ weaker = differences[column].gt(0).groupby(KEY).sum()
+ summary[name + "_tumor_stronger"] = stronger
+ summary[name + "_tumor_weaker"] = weaker
+ summary[name + "_top100_libraries"] = ranks.groupby(KEY)[
+ ("base" if name == "baseline" else "next") + "_top100"
+ ].sum()
+ evaluable = summary["pairs"].ge(5)
+ p = [
+ binomtest(s, s + w).pvalue if s + w else 1.0
+ for s, w in zip(stronger[evaluable], weaker[evaluable], strict=True)
+ ]
+ summary.loc[evaluable, name + "_sign_test_bh"] = false_discovery_control(p)
+ summary.to_csv(
+ ROOT / "scc-paired-rank-summary.csv.gz", compression={"method": "gzip", "mtime": 0}
+ )
+ differences.to_csv(
+ ROOT / "scc-paired-rank-donors.csv.gz", compression={"method": "gzip", "mtime": 0}
+ )
+ subset = summary[summary.pairs >= 5]
+ flip = subset.baseline_median_delta_pp * subset.candidate_median_delta_pp < 0
+ print("All measured combinations", ranks.groupby(KEY).ngroups, "paired", len(summary))
+ print("At least five pairs", len(subset), "median changes sign", flip.sum())
+ print(
+ "Median zero in either version",
+ ((subset.baseline_median_delta_pp == 0) | (subset.candidate_median_delta_pp == 0)).sum(),
+ )
+ for name in ("baseline", "candidate"):
+ print(name, "minimum adjusted p", subset[name + "_sign_test_bh"].min())
+ print(
+ subset[flip]
+ .assign(
+ shift=np.minimum(
+ subset.baseline_median_delta_pp.abs(), subset.candidate_median_delta_pp.abs()
+ )
+ )
+ .nlargest(5, "shift")
+ .to_string()
+ )
+ plot_donor_changes(differences)
+
+
+def plot_donor_changes(differences: pd.DataFrame) -> None:
+ import matplotlib.pyplot as plt
+ from matplotlib.lines import Line2D
+
+ panels = pd.read_csv(ROOT / "scc-publication-donor-differences.csv").set_index("donor")
+ plt.rcParams["svg.fonttype"] = "none"
+ fig, axes = plt.subplots(2, 2, figsize=(8.4, 6.3), layout="constrained")
+ rank_columns = ["baseline_tumor_minus_normal", "candidate_tumor_minus_normal"]
+ examples = [
+ (
+ "A TGFB1 → ENG",
+ "Keratinocyte to fibroblast",
+ panels.loc[panels.panel == "communication.liana", rank_columns],
+ "Magnitude score",
+ ),
+ (
+ "B Tensor factor 1",
+ "Matched sample loadings",
+ panels.loc[panels.panel == "communication.tensor", rank_columns],
+ "Factor loading",
+ ),
+ (
+ "C COL18A1 → ITGB5",
+ "Fibroblast to fibroblast",
+ differences.loc[("Fibroblast", "Fibroblast", "COL18A1", "ITGB5")],
+ "Rank percentile",
+ ),
+ (
+ "D NAMPT → ITGA5–ITGB1",
+ "Keratinocyte to fibroblast",
+ differences.loc[("KC_other", "Fibroblast", "NAMPT", "ITGA5_ITGB1")],
+ "Rank percentile",
+ ),
+ ]
+ colors = ("#376783", "#AD4B32")
+ for ax, (title, subtitle, frame, units) in zip(axes.flat, examples, strict=True):
+ data = frame.loc[sorted(frame.index, key=lambda p: int(p[1:]))]
+ positions = np.arange(len(data))
+ ax.axvline(0, color="#959595", lw=0.8, ls="--")
+ ax.hlines(positions, data.iloc[:, 0], data.iloc[:, 1], color="#8a8a8a", lw=1.1)
+ ax.scatter(
+ data.iloc[:, 0], positions, facecolor="white", edgecolor=colors[0], s=48, zorder=3
+ )
+ ax.scatter(data.iloc[:, 1], positions, color=colors[1], s=15, zorder=4)
+ ax.set_yticks(positions, data.index)
+ ax.set_ylim(len(data) - 0.45, -0.65)
+ ax.set_title(title + "\n" + subtitle, loc="left", fontsize=10)
+ ax.set_xlabel(units + " difference (tumor − normal)", fontsize=9)
+ ax.spines[["top", "right", "left"]].set_visible(False)
+ ax.tick_params(labelsize=8)
+ handles = [
+ Line2D(
+ [],
+ [],
+ ls="",
+ marker="o",
+ markerfacecolor="white",
+ markeredgecolor=colors[0],
+ label="LIANA 1.10.0",
+ ),
+ Line2D([], [], ls="", marker="o", color=colors[1], label="LIANA 2.1.1"),
+ ]
+ fig.legend(handles=handles, loc="outside lower center", ncol=2, frameon=False)
+ for suffix in ("png", "svg"):
+ fig.savefig(ROOT / ("scc-paired-rank-changes." + suffix), dpi=180)
+ plt.close(fig)
+
+
+if __name__ == "__main__":
+ main()
diff --git a/scripts/verify_gpu.py b/scripts/verify_gpu.py
index 09d53725..98557cb3 100644
--- a/scripts/verify_gpu.py
+++ b/scripts/verify_gpu.py
@@ -14,6 +14,12 @@
def verify_rapids() -> int:
try:
+ from cellquorum.backends.rapids_worker import worker_python, worker_status
+
+ if worker := worker_python():
+ print("OK: isolated RAPIDS kernel", worker_status(worker))
+ return 0
+
import cupy as cp
import rapids_singlecell
diff --git a/scripts/verify_reference_study.py b/scripts/verify_reference_study.py
index 6513f344..793a9050 100644
--- a/scripts/verify_reference_study.py
+++ b/scripts/verify_reference_study.py
@@ -10,6 +10,8 @@
import pandas as pd
from scipy.stats import spearmanr
+from cellquorum.core.stage import StageExecutionRecord
+
def require(condition: bool, message: str) -> None:
if not condition:
@@ -348,9 +350,11 @@ def check_integration(run: Path, data: Path) -> dict:
for domain in ["bio", "batch"]
for key, score in domains[domain].items()
]
- table = pd.DataFrame(rows, columns=["embedding", "category", "metric", "value"])
+ columns = ["embedding", "category", "metric"]
+ expected = pd.DataFrame(rows, columns=[*columns, "value"]).set_index(columns).sort_index()
+ table = pd.read_csv(run / "reports/integration_quality.csv").set_index(columns).sort_index()
+ pd.testing.assert_frame_equal(table, expected, check_exact=False, rtol=1e-12, atol=1e-14)
require(np.isfinite(table.value).all(), "Integration metric missing.")
- table.to_csv(run / "reports/integration_quality.csv", index=False)
raw, harmony = scores["X_pca"], scores["X_pca_harmony"]
require(
harmony["batch"]["ilisi"] > raw["batch"]["ilisi"]
@@ -603,6 +607,25 @@ def check_extended_lineage(run: Path, data: Path) -> dict:
"median_potency": medians.to_dict(),
"fates": fates.shape[1],
}
+ publication = run / "reports/publication"
+ for stem in ("lineage.palantir", "lineage"):
+ for extension in ("pdf", "svg"):
+ page = publication / f"{stem}.{extension}"
+ require(page.is_file() and page.stat().st_size > 0, f"Missing figure: {page.name}")
+ panel = pd.read_csv(publication / "lineage.palantir.csv").set_index("group")
+ expected = obj.obs.groupby("clusters", observed=True).palantir_pseudotime.agg(
+ ["count", "median"]
+ )
+ require(
+ panel.index.is_unique and set(expected.index) == set(panel.index),
+ "Palantir figure populations differ from the analysis object.",
+ )
+ expected = expected.loc[panel.index]
+ require(
+ np.array_equal(panel.n_cells, expected["count"])
+ and np.allclose(panel.value, expected["median"], rtol=0, atol=1e-12),
+ "Palantir figure cell counts or medians differ from the analysis object.",
+ )
settings = json.loads((run / "provenance/resolved_config.json").read_text())[
"trajectory_viz"
]
@@ -817,7 +840,11 @@ def main() -> None:
records = json.loads((run / "provenance/stage_execution_records.json").read_text())
plan = json.loads((run / "provenance/pipeline_plan.json").read_text())
enabled = {s["name"] for s in plan["stages"] if s["enabled"]}
- complete = {r["stage_name"] for r in records if r["status"] == "success"}
+ complete = {
+ record.stage_name
+ for record in map(StageExecutionRecord.from_dict, records)
+ if record.status == "success" or record.reused_from_checkpoint
+ }
require(
enabled <= complete,
f"Enabled stages without successful execution: {sorted(enabled - complete)}",
diff --git a/src/cellquorum/backends/celloracle_scripts/celloracle_ko.py b/src/cellquorum/backends/celloracle_scripts/celloracle_ko.py
index 36724ab6..e61952c2 100644
--- a/src/cellquorum/backends/celloracle_scripts/celloracle_ko.py
+++ b/src/cellquorum/backends/celloracle_scripts/celloracle_ko.py
@@ -233,7 +233,11 @@ def _run_celloracle(args: argparse.Namespace) -> None: # pragma: no cover
if int(args.max_cells) > 0 and adata.n_obs > int(args.max_cells):
cond_key = args.condition_key if args.condition_key in adata.obs.columns else None
keep = _stratified_subsample(
- adata.obs, cluster_key, cond_key, int(args.max_cells), int(args.seed)
+ adata.obs if isinstance(adata.obs, pd.DataFrame) else adata.obs.to_memory(),
+ cluster_key,
+ cond_key,
+ int(args.max_cells),
+ int(args.seed),
)
print(
f"[celloracle] Subsampling {adata.n_obs} -> {len(keep)} cells "
diff --git a/src/cellquorum/backends/compute.py b/src/cellquorum/backends/compute.py
index 411aabe9..c533b549 100644
--- a/src/cellquorum/backends/compute.py
+++ b/src/cellquorum/backends/compute.py
@@ -33,6 +33,14 @@ def gpu_compute_available() -> bool:
Whether GPU compute (rapids-singlecell) is usable in this process.
"""
+ from cellquorum.backends.rapids_worker import worker_python, worker_status
+
+ if worker := worker_python():
+ try:
+ return bool(worker_status(worker)["cuda_visible"])
+ except (OSError, RuntimeError, ValueError):
+ return False
+
global _GPU_AVAILABLE
if _GPU_AVAILABLE is not None:
return _GPU_AVAILABLE
diff --git a/src/cellquorum/backends/metacells_scripts/run_metacell2.py b/src/cellquorum/backends/metacells_scripts/run_metacell2.py
index c556cd9b..1433492a 100644
--- a/src/cellquorum/backends/metacells_scripts/run_metacell2.py
+++ b/src/cellquorum/backends/metacells_scripts/run_metacell2.py
@@ -132,6 +132,7 @@ def main(argv: list[str]) -> int:
# ---- Run the MetaCell2 pipeline ---- #
import numpy as np
+ from scipy import sparse
if layer != "X":
full.X = full.layers[layer]
@@ -140,8 +141,13 @@ def main(argv: list[str]) -> int:
# int64/int32 counts (the common on-disk dtype) fail that assertion with
# no informative message, deep inside the pipeline. Synthetic data built
# with float32 by construction never exercised this.
- if full.X.dtype != np.float32:
- full.X = full.X.astype(np.float32)
+ counts = full.X
+ if not isinstance(counts, np.ndarray | sparse.spmatrix | sparse.sparray):
+ _write_empty(out_dir, "MetaCell2 requires a dense or sparse count matrix")
+ return 0
+ if counts.dtype != np.float32:
+ counts = counts.astype(np.float32)
+ full.X = counts
try:
mc.ut.set_name(full, "full")
@@ -182,7 +188,7 @@ def main(argv: list[str]) -> int:
metacell_adata = mc.pl.collect_metacells(clean, name="metacells", random_seed=seed)
# ---- Build the per-cell assignment table, including Outliers ---- #
- n_umis = np.asarray(full.X.sum(axis=1)).ravel()
+ n_umis = np.asarray(counts.sum(axis=1)).ravel()
groups = full.obs[group_by].astype(str).to_numpy()
cell_ids = full.obs_names.to_numpy()
diff --git a/src/cellquorum/backends/pyucell_scripts/ucell_score.py b/src/cellquorum/backends/pyucell_scripts/ucell_score.py
index a2d8105a..600c9b37 100644
--- a/src/cellquorum/backends/pyucell_scripts/ucell_score.py
+++ b/src/cellquorum/backends/pyucell_scripts/ucell_score.py
@@ -104,7 +104,10 @@ def main(argv: list[str] | None = None) -> int:
return 0
out_csv.parent.mkdir(parents=True, exist_ok=True)
- scores = adata.obs[score_cols].copy()
+ import pandas as pd
+
+ obs = adata.obs if isinstance(adata.obs, pd.DataFrame) else adata.obs.to_memory()
+ scores = obs[score_cols].copy()
scores.insert(0, "cell", adata.obs_names.to_numpy())
scores.to_csv(out_csv, index=False)
print(f"[pyucell] wrote {len(score_cols)} signature score(s) -> {out_csv}")
diff --git a/src/cellquorum/backends/rapids.py b/src/cellquorum/backends/rapids.py
index 4446155c..07d10b7a 100644
--- a/src/cellquorum/backends/rapids.py
+++ b/src/cellquorum/backends/rapids.py
@@ -84,6 +84,29 @@ def status(self) -> BackendStatus:
warnings, and backend details.
"""
+ from cellquorum.backends.rapids_worker import worker_python, worker_status
+
+ if worker := worker_python():
+ try:
+ worker_details = worker_status(worker)
+ except (OSError, RuntimeError, ValueError) as exc:
+ return BackendStatus(
+ name=self.name,
+ kind=self.kind,
+ available=False,
+ missing=["RAPIDS worker"],
+ warnings=[str(exc)],
+ details={"python": worker},
+ requirements=self.requirements(),
+ )
+ return BackendStatus(
+ name=self.name,
+ kind=self.kind,
+ available=True,
+ details={**worker_details, "python": worker},
+ requirements=self.requirements(),
+ )
+
missing: list[str] = []
warnings: list[str] = []
diff --git a/src/cellquorum/backends/rapids_worker.py b/src/cellquorum/backends/rapids_worker.py
new file mode 100644
index 00000000..c7cf9838
--- /dev/null
+++ b/src/cellquorum/backends/rapids_worker.py
@@ -0,0 +1,189 @@
+"""Execute RAPIDS kernels in-process or with an isolated pandas/cuDF install.
+
+The external interpreter runs this file directly and only needs the worker requirements.
+The controller exchanges indexed H5AD files and only imports the requested results.
+"""
+
+from __future__ import annotations
+
+import json
+import os
+import subprocess
+import sys
+from functools import lru_cache
+from pathlib import Path
+from tempfile import TemporaryDirectory
+from typing import TYPE_CHECKING, Any
+
+if TYPE_CHECKING:
+ from anndata import AnnData
+
+
+def worker_python() -> str | None:
+ return os.environ.get("CELLQUORUM_RAPIDS_PYTHON") or None
+
+
+def _invoke(python: str, *args: str) -> str:
+ process = subprocess.run(
+ [python, "-I", str(Path(__file__).resolve()), *args], capture_output=True, text=True
+ )
+ if process.returncode:
+ reason = process.stderr.strip() or process.stdout.strip()
+ raise RuntimeError(f"RAPIDS worker failed ({process.returncode}): {reason[-2000:]}")
+ return process.stdout
+
+
+@lru_cache(maxsize=4)
+def worker_status(python: str) -> dict[str, Any]:
+ return json.loads(_invoke(python, "probe").strip().splitlines()[-1])
+
+
+def _probe() -> dict[str, Any]:
+ from importlib.metadata import version
+
+ import cupy
+ import rapids_singlecell # noqa: F401
+
+ if (cupy.arange(8) + 1).sum().item() != 36:
+ raise RuntimeError("CUDA reduction returned an incorrect result")
+ device = cupy.cuda.runtime.getDeviceProperties(cupy.cuda.Device().id)
+ return {
+ "cuda_visible": True,
+ "cuda_device_count": cupy.cuda.runtime.getDeviceCount(),
+ "device": device["name"].decode(),
+ "versions": {
+ p: version(p) for p in ("rapids-singlecell-cu12", "cudf-cu12", "pandas", "anndata")
+ },
+ }
+
+
+def _apply(adata: AnnData, operation: str, options: dict[str, Any]) -> None:
+ import rapids_singlecell as rsc
+
+ if operation == "pca":
+ layer = options["layer"]
+ rsc.get.anndata_to_GPU(adata, layer=layer)
+ try:
+ rsc.pp.pca(adata, **options)
+ finally:
+ rsc.get.anndata_to_CPU(adata, layer=layer)
+ elif operation == "cluster":
+ rep = options["use_rep"]
+ if rep == "X":
+ rsc.get.anndata_to_GPU(adata)
+ try:
+ rsc.pp.neighbors(
+ adata, **{k: options[k] for k in ("n_neighbors", "use_rep", "random_state")}
+ )
+ rsc.tl.leiden(
+ adata, **{k: options[k] for k in ("resolution", "random_state", "key_added")}
+ )
+ finally:
+ if rep == "X":
+ rsc.get.anndata_to_CPU(adata)
+ elif operation == "harmony":
+ rsc.pp.harmony_integrate(adata, **options)
+ elif operation == "umap":
+ rsc.tl.umap(adata, **options)
+ else:
+ raise ValueError(f"Unknown RAPIDS operation {operation!r}")
+
+
+def _indexed(adata: AnnData) -> AnnData:
+ import anndata as ad
+ import pandas as pd
+
+ return ad.AnnData(
+ obs=pd.DataFrame(index=adata.obs_names), var=pd.DataFrame(index=adata.var_names)
+ )
+
+
+def _inputs(adata: AnnData, operation: str, options: dict[str, Any]) -> AnnData:
+ data = _indexed(adata)
+ if operation == "pca":
+ layer = options["layer"]
+ if layer is None:
+ data.X = adata.X
+ else:
+ data.layers[layer] = adata.layers[layer]
+ if mask := options["mask_var"]:
+ data.var[mask] = adata.var[mask]
+ elif operation == "cluster":
+ rep = options["use_rep"]
+ if rep == "X":
+ data.X = adata.X
+ else:
+ data.obsm[rep] = adata.obsm[rep]
+ elif operation == "harmony":
+ rep, key = options["basis"], options["key"]
+ data.obsm[rep] = adata.obsm[rep]
+ data.obs[key] = adata.obs[key]
+ elif operation == "umap":
+ neighbors = adata.uns["neighbors"]
+ data.uns["neighbors"] = neighbors
+ for slot in ("distances", "connectivities"):
+ key = neighbors.get(slot + "_key", slot)
+ if key in adata.obsp:
+ data.obsp[key] = adata.obsp[key]
+ rep = neighbors.get("params", {}).get("use_rep")
+ if rep and rep != "X":
+ data.obsm[rep] = adata.obsm[rep]
+ elif rep is None and "X_pca" in adata.obsm:
+ data.obsm["X_pca"] = adata.obsm["X_pca"]
+ else:
+ data.X = adata.X
+ return data
+
+
+def _copy_results(source: AnnData, dest: AnnData, operation: str, options: dict[str, Any]) -> None:
+ if not source.obs_names.equals(dest.obs_names) or not source.var_names.equals(dest.var_names):
+ raise ValueError("RAPIDS result cell or gene identifiers differ from input")
+ if operation == "pca":
+ dest.obsm["X_pca"] = source.obsm["X_pca"]
+ dest.varm["PCs"] = source.varm["PCs"]
+ dest.uns["pca"] = source.uns["pca"]
+ elif operation == "cluster":
+ key = options["key_added"]
+ dest.obs[key] = source.obs[key]
+ dest.uns[key] = source.uns[key]
+ dest.uns["neighbors"] = source.uns["neighbors"]
+ for slot in ("connectivities", "distances"):
+ graph = source.uns["neighbors"].get(slot + "_key", slot)
+ dest.obsp[graph] = source.obsp[graph]
+ elif operation == "harmony":
+ key = options["adjusted_basis"]
+ dest.obsm[key] = source.obsm[key]
+ elif operation == "umap":
+ dest.obsm["X_umap"] = source.obsm["X_umap"]
+ dest.uns["umap"] = source.uns["umap"]
+
+
+def run_rapids(adata: AnnData, operation: str, **options: Any) -> None:
+ python = worker_python()
+ if not python:
+ _apply(adata, operation, options)
+ return
+ import anndata as ad
+
+ with TemporaryDirectory(prefix="cellquorum-rapids-") as scratch:
+ folder = Path(scratch)
+ _inputs(adata, operation, options).write_h5ad(folder / "input.h5ad")
+ _invoke(python, operation, scratch, json.dumps(options))
+ result = ad.read_h5ad(folder / "output.h5ad")
+ _copy_results(result, adata, operation, options)
+
+
+if __name__ == "__main__":
+ if sys.argv[1] == "probe":
+ print(json.dumps(_probe()))
+ else:
+ import anndata as ad
+
+ operation, directory, raw_options = sys.argv[1:]
+ folder = Path(directory)
+ options = json.loads(raw_options)
+ data = ad.read_h5ad(folder / "input.h5ad")
+ _apply(data, operation, options)
+ result = _indexed(data)
+ _copy_results(data, result, operation, options)
+ result.write_h5ad(folder / "output.h5ad")
diff --git a/src/cellquorum/cli/study.py b/src/cellquorum/cli/study.py
index 9fc782ae..01b3ffe5 100644
--- a/src/cellquorum/cli/study.py
+++ b/src/cellquorum/cli/study.py
@@ -5,6 +5,7 @@
import os
import shlex
import tempfile
+from collections.abc import Sized
from pathlib import Path
from typing import Any
@@ -18,6 +19,7 @@
from cellquorum.config.design import DesignConfig, validate_design_against_obs
from cellquorum.config.models import CellQuorumConfig, StageSelectionConfig
+from cellquorum.core.anndata_access import obs_frame
from cellquorum.core.labels import as_label_strings
from cellquorum.stages.qc.config import QCFloorConfig
@@ -45,14 +47,15 @@ def _metadata(path: Path) -> tuple[pd.DataFrame, int, list[str]]:
if isinstance(handle["obs"], h5py.Group) and "_index" in handle["var"].attrs:
obs = read_elem(handle["obs"])
var = handle["var"]
- gene_names = read_elem(var[var.attrs["_index"]])
- return obs, len(gene_names), layers
+ index = read_elem(var[var.attrs["_index"]])
+ if isinstance(obs, pd.DataFrame) and isinstance(index, Sized):
+ return obs, len(index), layers
# Older files store categorical codes in obs and their labels in uns. AnnData's
# backwards-compatible reader rejoins them before we show any values to the user.
data = ad.read_h5ad(path, backed="r")
try:
- return data.obs, data.n_vars, layers
+ return obs_frame(data), data.n_vars, layers
finally:
data.file.close()
diff --git a/src/cellquorum/config/cohort.py b/src/cellquorum/config/cohort.py
index 8190c7ac..258b05c2 100644
--- a/src/cellquorum/config/cohort.py
+++ b/src/cellquorum/config/cohort.py
@@ -15,6 +15,8 @@
from __future__ import annotations
+from collections.abc import Mapping
+
from cellquorum.config.base import StrictBaseModel
@@ -76,12 +78,28 @@ def resolve_cohort_key(
The cohort value when set, else ``stage_value``.
"""
- cohort = getattr(config, "cohort", None)
- if cohort is not None:
- cohort_value = getattr(cohort, attr, None)
- if cohort_value:
- return cohort_value
- return stage_value
+ cohort = (
+ config.get("cohort") if isinstance(config, Mapping) else getattr(config, "cohort", None)
+ )
+ value = cohort.get(attr) if isinstance(cohort, Mapping) else getattr(cohort, attr, None)
+ return value or stage_value
+
+
+def resolve_design_columns(
+ config: object, *, donor_key: str | None, condition_key: str | None
+) -> tuple[str, str]:
+ """Resolve cohort and stage column names, with the study design as their fallback."""
+ design = (
+ config.get("design") if isinstance(config, Mapping) else getattr(config, "design", None)
+ )
+ donor = resolve_cohort_key(config, attr="donor_key", stage_value=donor_key)
+ condition = resolve_cohort_key(config, attr="condition_key", stage_value=condition_key)
+ if isinstance(design, Mapping):
+ design_donor, design_condition = design.get("donor_col"), design.get("condition_col")
+ else:
+ design_donor = getattr(design, "donor_col", None)
+ design_condition = getattr(design, "condition_col", None)
+ return donor or design_donor or "patient_id", condition or design_condition or "condition"
def validate_cohort_against_obs(obs_columns: list[str], *, cohort: CohortConfig) -> list[str]:
@@ -117,5 +135,6 @@ def validate_cohort_against_obs(obs_columns: list[str], *, cohort: CohortConfig)
"CohortConfig",
"FocusLineageConfig",
"resolve_cohort_key",
+ "resolve_design_columns",
"validate_cohort_against_obs",
]
diff --git a/src/cellquorum/core/anndata_access.py b/src/cellquorum/core/anndata_access.py
new file mode 100644
index 00000000..f2525be0
--- /dev/null
+++ b/src/cellquorum/core/anndata_access.py
@@ -0,0 +1,61 @@
+"""Access AnnData annotations and expression arrays for pandas and SciPy operations."""
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+import numpy as np
+import pandas as pd
+from scipy import sparse
+
+if TYPE_CHECKING:
+ import anndata as ad
+
+
+def resolve_use_rep(adata: ad.AnnData, configured: str | None, fallback: list[str]) -> str | None:
+ """Select the requested embedding or the first available fallback in the given order."""
+ if configured and configured in adata.obsm:
+ return configured
+ return next((name for name in fallback if name in adata.obsm), None)
+
+
+def obs_frame(adata: ad.AnnData) -> pd.DataFrame:
+ """Return the object's writable cell metadata, materializing only that axis if lazy."""
+ if not isinstance(adata.obs, pd.DataFrame):
+ adata.obs = adata.obs.to_memory()
+ return adata.obs
+
+
+def var_frame(adata: ad.AnnData) -> pd.DataFrame:
+ """Return the object's writable feature metadata, materializing only that axis if lazy."""
+ if not isinstance(adata.var, pd.DataFrame):
+ adata.var = adata.var.to_memory()
+ return adata.var
+
+
+def memory_matrix(value: object) -> np.ndarray | sparse.spmatrix | sparse.sparray:
+ """Obtain a CPU expression matrix for a NumPy/SciPy consumer, retaining sparsity."""
+ if value is None:
+ raise ValueError("The requested expression matrix is absent")
+ to_memory = getattr(value, "to_memory", None)
+ if callable(to_memory):
+ value = to_memory()
+ if isinstance(value, sparse.spmatrix | sparse.sparray):
+ return value
+ matrix = np.asarray(value)
+ if matrix.ndim != 2:
+ raise ValueError("Expected a two-dimensional expression matrix")
+ return matrix
+
+
+def copy_x_to_layer(adata: ad.AnnData, name: str) -> None:
+ """Preserve the original array backend when copying X into a named layer."""
+ matrix = adata.X
+ copy = getattr(matrix, "copy", None)
+ adata.layers[name] = copy() if callable(copy) else memory_matrix(matrix).copy()
+
+
+def dense_matrix(value: object) -> np.ndarray:
+ """Materialize the selected matrix for a consumer requiring a dense array."""
+ matrix = memory_matrix(value)
+ return matrix if isinstance(matrix, np.ndarray) else matrix.toarray()
diff --git a/src/cellquorum/core/checkpoint.py b/src/cellquorum/core/checkpoint.py
index 725e765f..3f8f0f8d 100644
--- a/src/cellquorum/core/checkpoint.py
+++ b/src/cellquorum/core/checkpoint.py
@@ -38,7 +38,6 @@
from __future__ import annotations
-import contextlib
import json
import shutil
from dataclasses import dataclass
@@ -46,7 +45,8 @@
from pathlib import Path
from typing import TYPE_CHECKING
-from cellquorum.core.fingerprint import FINGERPRINT_SCHEMA_VERSION
+from cellquorum.core.fingerprint import FINGERPRINT_SCHEMA_VERSION, artifact_file_checksums
+from cellquorum.core.stage import StageExecutionRecord
if TYPE_CHECKING:
import anndata as ad
@@ -92,6 +92,9 @@ class CheckpointRecord:
# settings genuinely changed, and the staleness message accuses the wrong thing.
fingerprint_schema_version: int = FINGERPRINT_SCHEMA_VERSION
+ execution_records: list[StageExecutionRecord] | None = None
+ artifact_checksums: dict[str, str] | None = None
+
@property
def object_path(self) -> Path:
return self.path / _OBJECT_NAME
@@ -132,6 +135,7 @@ def write_checkpoint(
order: int,
input_fingerprint: str | None,
upstream_fingerprint: str | None = None,
+ execution_records: list[StageExecutionRecord] | None = None,
) -> CheckpointRecord | None:
"""Write ``adata`` as ``stage``'s checkpoint. None when there is nothing to write.
@@ -150,18 +154,19 @@ def write_checkpoint(
# holding X and obs and nothing else, which is unusable as a resume point but
# looks like a checkpoint on disk.
from cellquorum.core.h5ad_io import H5adWriteError, write_h5ad
+ from cellquorum.core.provenance import to_json_safe
+ marker = target / _SIDECAR_NAME
try:
+ # A killed writer must never pair a new object with the previous receipt.
+ marker.unlink(missing_ok=True)
write_h5ad(adata, object_path)
- except H5adWriteError as exc:
- # The writer has already removed any object left at that path, so drop the
- # sidecar too: discovery keys on the sidecar, and one without its object is
- # a checkpoint that exists as far as `--from-stage` is concerned and then
- # fails on read. Half a checkpoint is worse than none.
- with contextlib.suppress(OSError):
- (target / _SIDECAR_NAME).unlink(missing_ok=True)
+ except (OSError, H5adWriteError) as exc:
raise CheckpointError(f"could not write checkpoint for '{stage}': {exc}") from exc
+ checksums = artifact_file_checksums(
+ a.path for row in execution_records or [] for a in row.output_artifacts
+ )
record = CheckpointRecord(
stage=stage,
order=int(order),
@@ -170,24 +175,35 @@ def write_checkpoint(
n_vars=int(adata.n_vars),
path=target,
upstream_fingerprint=upstream_fingerprint,
+ execution_records=execution_records,
+ artifact_checksums=checksums,
)
- (target / _SIDECAR_NAME).write_text(
- json.dumps(
- {
- "schema_version": CHECKPOINT_SCHEMA_VERSION,
- "fingerprint_schema_version": record.fingerprint_schema_version,
- "stage": record.stage,
- "order": record.order,
- "input_fingerprint": record.input_fingerprint,
- "upstream_fingerprint": record.upstream_fingerprint,
- "n_obs": record.n_obs,
- "n_vars": record.n_vars,
- "object": _OBJECT_NAME,
- },
- indent=2,
+ pending = marker.with_suffix(".tmp")
+ try:
+ pending.write_text(
+ json.dumps(
+ {
+ "schema_version": CHECKPOINT_SCHEMA_VERSION,
+ "fingerprint_schema_version": record.fingerprint_schema_version,
+ "stage": record.stage,
+ "order": record.order,
+ "input_fingerprint": record.input_fingerprint,
+ "upstream_fingerprint": record.upstream_fingerprint,
+ "n_obs": record.n_obs,
+ "n_vars": record.n_vars,
+ "object": _OBJECT_NAME,
+ "artifact_checksums": record.artifact_checksums,
+ "execution_records": None
+ if execution_records is None
+ else to_json_safe([row.to_dict() for row in execution_records]),
+ },
+ indent=2,
+ )
+ + "\n"
)
- + "\n"
- )
+ pending.replace(marker)
+ except (ValueError, TypeError, OSError) as exc:
+ raise CheckpointError(f"could not publish checkpoint for '{stage}': {exc}") from exc
return record
@@ -199,22 +215,27 @@ def read_checkpoint_record(paths: object, stage: str) -> CheckpointRecord | None
return None
try:
payload = json.loads(sidecar.read_text())
- except Exception: # noqa: BLE001 — a corrupt sidecar is simply "no checkpoint"
- return None
- if int(payload.get("schema_version", 0)) != CHECKPOINT_SCHEMA_VERSION:
+ if payload["schema_version"] != CHECKPOINT_SCHEMA_VERSION or payload["stage"] != stage:
+ return None
+ rows = payload.get("execution_records")
+ return CheckpointRecord(
+ stage=stage,
+ order=int(payload["order"]),
+ input_fingerprint=payload.get("input_fingerprint"),
+ n_obs=int(payload["n_obs"]),
+ n_vars=int(payload["n_vars"]),
+ path=target,
+ upstream_fingerprint=payload.get("upstream_fingerprint"),
+ fingerprint_schema_version=int(
+ payload.get("fingerprint_schema_version", _UNRECORDED_FINGERPRINT_SCHEMA)
+ ),
+ artifact_checksums=payload.get("artifact_checksums"),
+ execution_records=None
+ if rows is None
+ else [StageExecutionRecord.from_dict(r) for r in rows],
+ )
+ except (OSError, KeyError, TypeError, ValueError, AttributeError):
return None
- return CheckpointRecord(
- stage=str(payload.get("stage", stage)),
- order=int(payload.get("order", -1)),
- input_fingerprint=payload.get("input_fingerprint"),
- n_obs=int(payload.get("n_obs", -1)),
- n_vars=int(payload.get("n_vars", -1)),
- path=target,
- upstream_fingerprint=payload.get("upstream_fingerprint"),
- fingerprint_schema_version=int(
- payload.get("fingerprint_schema_version", _UNRECORDED_FINGERPRINT_SCHEMA)
- ),
- )
def available_checkpoints(paths: object) -> list[CheckpointRecord]:
@@ -250,7 +271,11 @@ def resolve_start_checkpoint(
f"unknown stage '{from_stage}'; expected one of {sorted(stage_order)}"
)
target_order = stage_order[from_stage]
- candidates = [r for r in available_checkpoints(paths) if r.order < target_order]
+ candidates = [
+ r
+ for r in available_checkpoints(paths)
+ if stage_order.get(r.stage) == r.order and r.order < target_order
+ ]
if not candidates:
existing = [r.stage for r in available_checkpoints(paths)]
pruned = [entry["stage"] for entry in _read_pruned_record(paths).get("pruned", [])]
diff --git a/src/cellquorum/core/context.py b/src/cellquorum/core/context.py
index c2408cd6..c878d49d 100644
--- a/src/cellquorum/core/context.py
+++ b/src/cellquorum/core/context.py
@@ -283,7 +283,7 @@ def require_manifest(self) -> pd.DataFrame:
return self.manifest
- def with_adata(self, adata: ad.AnnData) -> PipelineContext:
+ def with_adata(self, adata: ad.AnnData | None) -> PipelineContext:
"""
Return a shallow context copy with an updated AnnData object.
diff --git a/src/cellquorum/core/contracts/data_contract.py b/src/cellquorum/core/contracts/data_contract.py
index 44c6f2f2..d16b307b 100644
--- a/src/cellquorum/core/contracts/data_contract.py
+++ b/src/cellquorum/core/contracts/data_contract.py
@@ -49,6 +49,21 @@ class DataContract(StrictBaseModel):
# Guardrail: reject imputed data where statistics are not allowed.
forbid_imputed: bool = False
+ @classmethod
+ def counts_layer(cls, layer: str) -> DataContract:
+ """Require a named count layer and validate its provenance and values."""
+ return cls(required_layers=[layer], expression_layer=layer, expected_kind="counts")
+
+ @classmethod
+ def lognorm(cls, layer: str, *, required_obs: list[str] | None = None) -> DataContract:
+ """Validate normalized expression in X or a named layer alongside metadata requirements."""
+ return cls(
+ required_layers=[] if layer == "X" else [layer],
+ required_obs=required_obs or [],
+ expression_layer=layer,
+ expected_kind="lognorm",
+ )
+
def validate(self, adata: ad.AnnData) -> None: # type: ignore[override] # shadows pydantic's deprecated BaseModel.validate on purpose
"""
Validate an AnnData object against this contract.
diff --git a/src/cellquorum/core/executor.py b/src/cellquorum/core/executor.py
index 86241f89..d22372d7 100644
--- a/src/cellquorum/core/executor.py
+++ b/src/cellquorum/core/executor.py
@@ -262,6 +262,7 @@ def run(
context: PipelineContext,
plan: PipelinePlan,
reporter: RunReporter | None = None,
+ restored_records: list[StageExecutionRecord] | None = None,
) -> PipelineExecutionResult:
"""
Execute a pipeline plan against a pipeline context.
@@ -286,7 +287,7 @@ def run(
# Initialize mutable execution state.
current_context = context
- stage_execution_records: list[StageExecutionRecord] = []
+ stage_execution_records = list(restored_records or [])
# Execute or skip every planned stage in order, with progress tracking.
with reporter.progress(total=len(plan.stages)) as bar:
@@ -414,6 +415,7 @@ def run(
planned_stage=planned_stage,
stage_result=stage_result,
reporter=reporter,
+ execution_records=stage_execution_records,
)
# Stop after the requested stage, so a pipeline can be advanced one
@@ -441,6 +443,7 @@ def _maybe_write_checkpoint(
planned_stage: object,
stage_result: object,
reporter: object,
+ execution_records: list[StageExecutionRecord] | None = None,
) -> None:
"""Write this stage's checkpoint when configured. Never fails the run.
@@ -487,6 +490,7 @@ def _maybe_write_checkpoint(
order=order_of.get(stage_name, -1),
input_fingerprint=getattr(stage_result, "input_fingerprint", None),
upstream_fingerprint=upstream_fingerprint,
+ execution_records=execution_records,
)
except CheckpointError as exc:
if hasattr(reporter, "stage_note"):
@@ -605,19 +609,8 @@ def execute_planned_stage(
except Exception: # pragma: no cover - resume is best-effort
decision = None
- if decision is not None and decision.resume:
- ended_at = datetime.now(UTC)
- record = StageExecutionRecord.skipped(
- stage_name=planned_stage.name,
- reason=decision.reason,
- started_at_utc=started_at,
- ended_at_utc=ended_at,
- backend_used=None,
- details={"resumed": True},
- notes=[f"Resumed: {decision.reason}"],
- input_fingerprint=input_fingerprint,
- )
- return None, replace(record, output_artifacts=list(decision.artifacts))
+ if decision is not None and decision.resume and decision.original is not None:
+ return None, decision.original.as_reused(decision.reason)
# Execute the registered stage.
try:
diff --git a/src/cellquorum/core/fingerprint.py b/src/cellquorum/core/fingerprint.py
index c8589e41..2200bb7a 100644
--- a/src/cellquorum/core/fingerprint.py
+++ b/src/cellquorum/core/fingerprint.py
@@ -1,13 +1,8 @@
"""Deterministic stage input fingerprints for resume/caching.
-A stage's ``input_fingerprint`` answers one question on rerun: "are the inputs
-to this stage the same as last time?" It is intentionally *conservative and
-cheap* — it hashes the resolved stage config, the random seed, and the salient
-shape/identity of the input AnnData (dimensions, var names, and which
-layers/obsm/obs columns are present), but NOT the full expression matrices.
-Hashing whole matrices would be slow and is unnecessary: within a single
-project the upstream data is stable, and any config or dimensionality change
-already flips the fingerprint. This is the key that stage-level resume compares.
+Fingerprints cover resolved settings and the identity and structure of stage inputs.
+The read-only integration benchmark also hashes the actual labels, batches and
+embedding coordinates it scores, so equally shaped inputs can invalidate its cache.
"""
from __future__ import annotations
@@ -15,6 +10,7 @@
import csv
import glob
import json
+from collections.abc import Iterable
from hashlib import file_digest, sha256
from pathlib import Path
from typing import TYPE_CHECKING
@@ -37,7 +33,8 @@
# QC computed changed. Checkpoints written under v2 cannot be compared against v3
# fingerprints for the same reason v1-vs-v2 couldn't.
# v5 also includes source files listed in the library manifest.
-FINGERPRINT_SCHEMA_VERSION = 5
+# v6 binds checkpoints to compute settings and benchmarks to observed labels/coordinates.
+FINGERPRINT_SCHEMA_VERSION = 6
def _stable_json(payload: object) -> str:
@@ -149,6 +146,72 @@ def _adata_signature(adata: ad.AnnData | None) -> dict[str, object]:
}
+def _benchmark_inputs(adata: ad.AnnData, config: dict) -> dict:
+ """Hash the observations and numerical coordinates used by scib-benchmark."""
+ from importlib.metadata import PackageNotFoundError, version
+
+ import numpy as np
+ import pandas as pd
+
+ from cellquorum.core.anndata_access import dense_matrix
+
+ # Method-list entries can override the stage-level representations and labels.
+ variants = [
+ config,
+ *({**config, **m} for m in (config.get("methods") or []) if isinstance(m, dict)),
+ ]
+ columns: set[str] = set()
+ embeddings: set[str] = set()
+ for options in variants:
+ columns.update(
+ key
+ for key in (
+ options.get("batch_key", "batch"),
+ options.get("label_key", "cell_type"),
+ options.get("label_key_fallback"),
+ )
+ if isinstance(key, str) and key in adata.obs
+ )
+ embeddings.update(options.get("embeddings") or [])
+ embeddings.add(options.get("pre_embedding", "X_pca"))
+ obs = adata.obs[sorted(columns)]
+ if not isinstance(obs, pd.DataFrame):
+ obs = obs.to_memory()
+ labels = pd.util.hash_pandas_object(obs, index=True).to_numpy()
+ coordinates = {}
+ for name in sorted(embeddings.intersection(adata.obsm)):
+ matrix = adata.obsm[name]
+ digest = sha256(str((matrix.shape, str(getattr(matrix, "dtype", None)))).encode())
+ for start in range(0, adata.n_obs, 4096):
+ block = (
+ matrix.iloc[start : start + 4096]
+ if isinstance(matrix, pd.DataFrame)
+ else matrix[start : start + 4096]
+ )
+ values = dense_matrix(block)
+ digest.update(np.ascontiguousarray(values).tobytes())
+ coordinates[name] = digest.hexdigest()
+ versions: dict[str, str | None] = {}
+ for package in (
+ "scib-metrics",
+ "harmonypy",
+ "numpy",
+ "scikit-learn",
+ "jax",
+ "jaxlib",
+ "pynndescent",
+ ):
+ try:
+ versions[package] = version(package)
+ except PackageNotFoundError:
+ versions[package] = None
+ return {
+ "labels": sha256(labels.tobytes()).hexdigest(),
+ "embeddings": coordinates,
+ "versions": versions,
+ }
+
+
def compute_input_fingerprint(
*,
stage_name: str,
@@ -176,6 +239,8 @@ def compute_input_fingerprint(
"config": _without_outputs(_without_unset(stage_config or {})),
"adata": _adata_signature(adata),
}
+ if stage_name == "integration_benchmark" and adata is not None:
+ payload["benchmark_inputs"] = _benchmark_inputs(adata, stage_config or {})
return _hash_str(_stable_json(payload))
@@ -232,6 +297,10 @@ def compute_upstream_fingerprint(
"through_stage": through_stage,
"through_order": through_order,
"random_seed": run_block.get("random_seed"),
+ "compute": _without_unset(config.get("compute")),
+ "organism": (config.get("project") or {}).get("organism"),
+ "species_id": (config.get("project") or {}).get("species_id"),
+ "require_selected_methods": run_block.get("require_selected_methods", False),
"study_design": _without_unset(config.get("design")),
"cohort": _without_unset(config.get("cohort")),
"input": _input_signature(config.get("input")),
@@ -281,6 +350,15 @@ def resource_checksum(value: Path) -> str | None:
return _hash_str(_stable_json(files)) if files else None
+def artifact_file_checksums(paths: Iterable[Path]) -> dict[str, str]:
+ """Snapshot registered files so restored evidence can be verified before reuse."""
+ return {
+ str(path): digest
+ for path in set(paths)
+ if path.is_file() and (digest := resource_checksum(path)) is not None
+ }
+
+
def _library_resources(config: dict, upstream: list[str], enabled: dict) -> dict:
manifest = (config.get("paths") or {}).get("manifest")
active = {name for name in upstream if enabled.get(name)}
diff --git a/src/cellquorum/core/h5ad_io.py b/src/cellquorum/core/h5ad_io.py
index 49900623..392e5238 100644
--- a/src/cellquorum/core/h5ad_io.py
+++ b/src/cellquorum/core/h5ad_io.py
@@ -47,11 +47,15 @@
import numpy as np
import pandas as pd
+from cellquorum.core.anndata_access import obs_frame, var_frame
+
if TYPE_CHECKING:
import anndata as ad
+ from anndata.experimental.backed import Dataset2D
__all__ = [
"H5adWriteError",
+ "legacy_string_frame",
"sanitize_for_h5ad",
"write_h5ad",
]
@@ -61,6 +65,26 @@ class H5adWriteError(RuntimeError):
"""Raised when an AnnData could not be written to h5ad."""
+def legacy_string_frame(frame: pd.DataFrame | Dataset2D) -> pd.DataFrame:
+ """Encode string metadata for older H5AD readers and NumPy-based analysis workers."""
+ frame = frame.copy() if isinstance(frame, pd.DataFrame) else frame.to_memory(copy=True)
+ if isinstance(frame.index.dtype, pd.StringDtype):
+ frame.index = frame.index.astype(object)
+ for key, values in frame.items():
+ if isinstance(values.dtype, pd.StringDtype):
+ categories = pd.Index(values.dropna().unique(), dtype=object)
+ frame[key] = pd.Categorical(values, categories=categories)
+ elif isinstance(values.dtype, pd.CategoricalDtype) and isinstance(
+ values.cat.categories.dtype, pd.StringDtype
+ ):
+ frame[key] = pd.Categorical.from_codes(
+ values.cat.codes,
+ categories=values.cat.categories.astype(object),
+ ordered=values.cat.ordered,
+ )
+ return frame
+
+
def write_h5ad(adata: ad.AnnData, path: Path | str, *, sanitize: bool = True) -> list[str]:
"""Write ``adata`` to ``path``, atomically, returning notes on what was coerced.
@@ -156,12 +180,12 @@ def sanitize_for_h5ad(adata: ad.AnnData) -> list[str]:
"""
notes: list[str] = []
notes += _sanitize_uns_keys(adata.uns)
- notes += _sanitize_frame_columns(adata.obs, axis="obs")
- notes += _sanitize_frame_columns(adata.var, axis="var")
+ notes += _sanitize_frame_columns(obs_frame(adata), axis="obs")
+ notes += _sanitize_frame_columns(var_frame(adata), axis="var")
notes += _sanitize_mapping_keys(adata.obsm, name="obsm")
notes += _sanitize_mapping_keys(adata.varm, name="varm")
- notes += _coerce_unwritable_columns(adata.obs, axis="obs")
- notes += _coerce_unwritable_columns(adata.var, axis="var")
+ notes += _coerce_unwritable_columns(obs_frame(adata), axis="obs")
+ notes += _coerce_unwritable_columns(var_frame(adata), axis="var")
notes += _coerce_cellrank_lineage_arrays(adata)
notes += _jsonify_unwritable_uns(adata.uns)
return notes
@@ -269,8 +293,7 @@ def _coerce_unwritable_columns(frame: pd.DataFrame, *, axis: str) -> list[str]:
# when a projection matched no cells at all.
frame[col] = pd.Categorical([None] * len(series))
notes.append(
- f"{axis}['{col}']: entirely missing with no dtype; "
- "written as an empty categorical"
+ f"{axis}['{col}']: entirely missing with no dtype; written as an empty categorical"
)
continue
if all(isinstance(v, str) for v in present):
diff --git a/src/cellquorum/core/pipeline.py b/src/cellquorum/core/pipeline.py
index d3728bee..18520a29 100644
--- a/src/cellquorum/core/pipeline.py
+++ b/src/cellquorum/core/pipeline.py
@@ -7,7 +7,7 @@
import platform
import time
from collections.abc import Sequence
-from dataclasses import dataclass
+from dataclasses import dataclass, replace
from datetime import UTC, datetime
from pathlib import Path
from typing import Any
@@ -18,6 +18,7 @@
from cellquorum.backends.registry import BackendRegistry, build_default_backend_registry
from cellquorum.config.loader import load_config, save_resolved_config
from cellquorum.config.models import CellQuorumConfig
+from cellquorum.core.anndata_access import obs_frame
from cellquorum.core.artifacts import ArtifactManager
from cellquorum.core.context import PipelineContext, PipelinePaths
from cellquorum.core.executor import PipelineExecutionResult, PipelineExecutor
@@ -28,6 +29,7 @@
format_inherited_artifacts,
)
from cellquorum.core.planner import PipelinePlan, build_pipeline_plan
+from cellquorum.core.resume import completion_path, completion_payload, missing_outputs
from cellquorum.core.run_reporter import RunReporter
from cellquorum.core.stage import StageArtifact, StageExecutionRecord
from cellquorum.io import load_adata
@@ -480,70 +482,6 @@ def _stage_execution_records_dataframe(
)
-def _stage_completion_sidecar_payload(record: StageExecutionRecord) -> dict[str, object]:
- """
- Build the durable per-stage completion sidecar payload.
-
- The sidecar is intentionally redundant with ``stage_execution_records.json``.
- Resume logic should be able to inspect one stage's completion state without
- parsing the whole run log.
-
- Args:
- record: Successful stage execution record.
-
- Returns:
- JSON-safe completion sidecar payload.
- """
-
- # Build artifact existence summaries at sidecar-write time.
- artifact_status = [
- {
- **artifact.to_dict(),
- "exists": artifact.path.exists(),
- }
- for artifact in record.output_artifacts
- ]
-
- return {
- "schema_version": 1,
- "stage_name": record.stage_name,
- "status": record.status,
- "completed_at_utc": record.ended_at_utc.isoformat(),
- "duration_seconds": record.duration_seconds,
- "backend_used": record.backend_used,
- "method_version": record.method_version,
- "device": record.device,
- "input_fingerprint": record.input_fingerprint,
- "output_fingerprint": record.output_fingerprint,
- "checkpoint_path": None if record.checkpoint_path is None else str(record.checkpoint_path),
- "n_output_artifacts": len(record.output_artifacts),
- "output_artifacts": artifact_status,
- "metrics": dict(record.metrics),
- "warnings": list(record.warnings),
- "notes": list(record.notes),
- }
-
-
-def _stage_completion_sidecar_path(stage_name: str) -> Path:
- """
- Return a stable relative sidecar path for a stage.
-
- Args:
- stage_name: Stage name from the execution record.
-
- Returns:
- Relative path under the run root.
- """
-
- # Stage names are registry-controlled, but keep this safe for custom stages.
- safe_stage_name = "".join(
- char if char.isalnum() or char in {"_", "-"} else "_" for char in stage_name
- ).strip("_")
- if not safe_stage_name:
- safe_stage_name = "stage"
- return Path("provenance") / "stages" / safe_stage_name / "completion.json"
-
-
def _write_stage_completion_sidecars(
artifact_manager: ArtifactManager,
records: list[StageExecutionRecord],
@@ -561,9 +499,9 @@ def _write_stage_completion_sidecars(
if record.status != "success":
continue
artifact_manager.write_json(
- _stage_completion_sidecar_payload(record),
+ completion_payload(record),
name=f"stage_completion_{record.stage_name}",
- relative_path=_stage_completion_sidecar_path(record.stage_name),
+ relative_path=completion_path(Path("provenance"), record.stage_name),
description=f"Durable completion marker for stage '{record.stage_name}'.",
)
@@ -657,7 +595,7 @@ def _write_run_report_after_provenance(
config=config,
records=records,
artifact_manager=artifact_manager,
- obs=context.adata.obs if context.adata is not None else None,
+ obs=obs_frame(context.adata) if context.adata is not None else None,
)
for warning in warnings:
logger.warning("%s", warning)
@@ -916,28 +854,6 @@ def bootstrap_pipeline_run(
)
-def _restrict_plan_from_stage(plan: PipelinePlan, from_stage: str) -> PipelinePlan:
- """Return `plan` with stages before `from_stage` removed.
-
- Their results already live in the checkpoint being loaded, so re-running them
- would both waste time and risk producing a different object than the one the
- checkpoint recorded.
- """
- import dataclasses
-
- from cellquorum.core.stages import stage_order_map
-
- order_of = stage_order_map()
- start = order_of.get(from_stage)
- if start is None:
- return plan
- # PlannedStage records enablement, not order, so the cut point comes from the
- # catalog. Stages missing from the catalog are kept rather than dropped: losing
- # a stage silently is worse than running one extra.
- kept = [s for s in plan.stages if order_of.get(s.name, start) >= start]
- return dataclasses.replace(plan, stages=kept)
-
-
def _write_inherited_artifacts_report(
*, context: PipelineContext, artifacts: list[InheritedArtifact]
) -> Path:
@@ -1013,19 +929,20 @@ def _config_disabled_stage_names(
return sorted(disabled)
-def _context_from_checkpoint(
- context: PipelineContext, config: CellQuorumConfig, from_stage: str
-) -> PipelineContext:
+def _resume_from_checkpoint(
+ context: PipelineContext, config: CellQuorumConfig, from_stage: str, plan: PipelinePlan
+) -> tuple[PipelineContext, PipelinePlan, list[StageExecutionRecord]]:
"""Return ``context`` carrying the checkpoint object that ``from_stage`` resumes from.
Raises if no usable checkpoint exists, or if the one found was written under settings this
run is not using.
"""
from cellquorum.core.checkpoint import (
+ CheckpointError,
load_checkpoint,
resolve_start_checkpoint,
)
- from cellquorum.core.fingerprint import compute_upstream_fingerprint
+ from cellquorum.core.fingerprint import compute_upstream_fingerprint, resource_checksum
from cellquorum.core.stages import stage_order_map
stage_order = stage_order_map()
@@ -1036,7 +953,7 @@ def _context_from_checkpoint(
# this the run would resume happily and report success while its numbers came
# from a config nobody chose, and resolved_config.json would disagree with the
# object on disk.
- return context.with_adata(
+ context = context.with_adata(
load_checkpoint(
start_record,
expected_upstream_fingerprint=compute_upstream_fingerprint(
@@ -1047,6 +964,48 @@ def _context_from_checkpoint(
)
)
+ remaining = [
+ s
+ for s in plan.stages
+ if stage_order.get(s.name, start_record.order + 1) > start_record.order
+ ]
+ replay = [
+ s.name
+ for s in remaining
+ if s.enabled and stage_order.get(s.name, 0) < stage_order[from_stage]
+ ]
+ if replay:
+ plan.warnings.append(
+ f"Restored '{start_record.stage}'; replaying {', '.join(replay)} before '{from_stage}'."
+ )
+ records = []
+ for row in start_record.execution_records or []:
+ if row.status == "success" or row.reused_from_checkpoint:
+ promised = config.run.required_outputs.get(row.stage_name, [])
+ registered = {artifact.path for artifact in row.output_artifacts}
+ missing = [str(p) for p in registered if not p.exists()]
+ missing.extend(
+ str(p)
+ for p in missing_outputs(
+ tuple(context.paths.root / name for name in promised), registered
+ )
+ )
+ if missing:
+ raise CheckpointError(
+ f"Checkpoint '{start_record.stage}' is missing outputs for "
+ f"'{row.stage_name}': {', '.join(sorted(set(missing)))}. Re-run that stage."
+ )
+ for path in registered:
+ expected = (start_record.artifact_checksums or {}).get(str(path))
+ if expected is not None and resource_checksum(path) != expected:
+ raise CheckpointError(
+ f"Output '{path}' changed after the '{start_record.stage}' checkpoint. "
+ f"Re-run '{row.stage_name}' to restore its evidence."
+ )
+ row = row.as_reused(f"Restored from the '{start_record.stage}' AnnData checkpoint")
+ records.append(row)
+ return context, replace(plan, stages=remaining), records
+
def _runnable_stage_names(
plan: PipelinePlan, executor: PipelineExecutor, config_dict: dict[str, Any]
@@ -1191,13 +1150,29 @@ def execute_pipeline_run(
until_stage=until_stage,
)
+ if executor is not None and until_stage is not None:
+ resolved_executor = replace(executor, until_stage=until_stage)
+
# Resume from a checkpoint when asked to start mid-pipeline. Done here, after
# the plan exists, because the plan is what gets restricted — and it must fail
# loudly rather than silently starting from raw input, which would look like a
# resume while producing different numbers.
+ execution_plan = plan
+ restored_records: list[StageExecutionRecord] = []
+ if from_stage and until_stage:
+ names = [stage.name for stage in plan.stages]
+ if (
+ from_stage in names
+ and until_stage in names
+ and names.index(until_stage) < names.index(from_stage)
+ ):
+ raise ValueError(f"Stop stage '{until_stage}' precedes start stage '{from_stage}'.")
if from_stage:
- context = _context_from_checkpoint(context, config, from_stage)
- plan = _restrict_plan_from_stage(plan, from_stage)
+ context, execution_plan, restored_records = _resume_from_checkpoint(
+ context, config, from_stage, plan
+ )
+ if until_stage is not None and until_stage not in {s.name for s in execution_plan.stages}:
+ raise ValueError(f"Stop stage '{until_stage}' is outside the execution plan.")
reporter = RunReporter(
verbose=config.run.verbose,
@@ -1232,8 +1207,9 @@ def execute_pipeline_run(
# Execute registered stages from the plan with progress reporting.
execution_result = resolved_executor.run(
context=context,
- plan=plan,
+ plan=execution_plan,
reporter=reporter,
+ **({"restored_records": restored_records} if restored_records else {}),
)
# Measure elapsed wall-clock time.
diff --git a/src/cellquorum/core/publication.py b/src/cellquorum/core/publication.py
index f78f3a83..14985753 100644
--- a/src/cellquorum/core/publication.py
+++ b/src/cellquorum/core/publication.py
@@ -122,7 +122,11 @@ def _communication(
else settings.nichenet_sender
)
priorities = priorities[priorities.sender.isin(senders or [])]
- liana = liana.rename(
+ # LIANA 2 exports component genes alongside the complex identifiers.
+ # The join and plotted rank both describe the complete interaction.
+ liana = liana[
+ ["sample", "source", "target", "ligand_complex", "receptor_complex", "magnitude_rank"]
+ ].rename(
columns={
"source": "sender",
"target": "receiver",
diff --git a/src/cellquorum/core/resume.py b/src/cellquorum/core/resume.py
index 6b4bfd2f..628ddc2a 100644
--- a/src/cellquorum/core/resume.py
+++ b/src/cellquorum/core/resume.py
@@ -10,15 +10,9 @@
3. the stage is *side-effect only* — it does not transform the AnnData that
flows downstream.
-Condition (3) is the conservative guardrail for this first implementation: the
-completion sidecar stores artifact paths and fingerprints, but NOT the AnnData
-state at that point. A stage that mutates ``adata`` (QC filtering, PCA,
-clustering, integration) therefore cannot be resumed yet — skipping it would
-lose the transformed matrix for downstream stages. Read-only/diagnostic stages
-(annotation diagnostics, integration benchmark, population identity, report)
-pass ``adata`` through unchanged, so skipping them on a fingerprint match is
-safe. AnnData-checkpoint resume for compute-heavy stages is a documented
-follow-up.
+The integration benchmark returns metrics and leaves its input object unchanged.
+Annotation diagnostics write cell-level evidence into obs, and population identity
+writes its audit into uns. Both must run to reconstruct the downstream object.
"""
from __future__ import annotations
@@ -27,19 +21,13 @@
from dataclasses import dataclass
from pathlib import Path
-from cellquorum.core.stage import StageArtifact
+from cellquorum.core.fingerprint import artifact_file_checksums
+from cellquorum.core.stage import StageArtifact, StageExecutionRecord
# Stages whose ``run`` does not change the AnnData that flows downstream. Only
# these are eligible for resume until AnnData checkpointing exists. Keep this
# conservative: when in doubt, a stage is NOT listed and simply re-runs.
-RESUMABLE_STAGES: frozenset[str] = frozenset(
- {
- "annotation_diagnostics",
- "integration_benchmark",
- "population_identity",
- "report",
- }
-)
+RESUMABLE_STAGES: frozenset[str] = frozenset({"integration_benchmark"})
@dataclass(frozen=True)
@@ -51,10 +39,33 @@ class ResumeDecision:
# Human-readable reason (for notes/records) whether or not resuming.
reason: str
- artifacts: tuple[StageArtifact, ...] = ()
+ original: StageExecutionRecord | None = None
+
+ @property
+ def artifacts(self) -> tuple[StageArtifact, ...]:
+ return tuple(self.original.output_artifacts) if self.original else ()
+
+
+def completion_payload(record: StageExecutionRecord) -> dict:
+ """Store a restorable execution record alongside the artifact existence snapshot."""
+ payload = record.to_dict()
+ payload.update(
+ schema_version=1,
+ completed_at_utc=record.ended_at_utc.isoformat(),
+ n_output_artifacts=len(record.output_artifacts),
+ output_artifacts=[
+ {**artifact.to_dict(), "exists": artifact.path.exists()}
+ for artifact in record.output_artifacts
+ ],
+ )
+ if record.stage_name in RESUMABLE_STAGES:
+ payload["artifact_checksums"] = artifact_file_checksums(
+ a.path for a in record.output_artifacts
+ )
+ return payload
-def _completion_sidecar_path(provenance_dir: Path, stage_name: str) -> Path:
+def completion_path(provenance_dir: Path, stage_name: str) -> Path:
"""Return the completion.json path for a stage under a provenance dir."""
safe_stage_name = "".join(
@@ -95,7 +106,7 @@ def decide_stage_resume(
if input_fingerprint is None:
return ResumeDecision(False, "no input fingerprint; re-running")
- sidecar_path = _completion_sidecar_path(provenance_dir, stage_name)
+ sidecar_path = completion_path(provenance_dir, stage_name)
if not sidecar_path.exists():
return ResumeDecision(False, "no prior completion marker; re-running")
@@ -106,7 +117,7 @@ def decide_stage_resume(
return ResumeDecision(False, "unreadable completion marker; re-running")
# Only prior *successful* completions are reusable.
- if payload.get("status") != "success":
+ if not isinstance(payload, dict) or payload.get("status") != "success":
return ResumeDecision(False, "prior stage did not succeed; re-running")
# The fingerprint must match, or an input/config change invalidated it.
@@ -114,29 +125,25 @@ def decide_stage_resume(
if prior_fingerprint is None or prior_fingerprint != input_fingerprint:
return ResumeDecision(False, "input fingerprint changed; re-running")
- # Every recorded artifact must still exist, or outputs are incomplete.
- artifacts = []
- for artifact in payload.get("output_artifacts", []):
- artifact_path = artifact.get("path")
- if artifact_path is None or not Path(artifact_path).exists():
- return ResumeDecision(False, "a recorded artifact is missing; re-running")
- path = Path(artifact_path)
- artifacts.append(
- StageArtifact(
- name=artifact.get("name", path.name),
- path=path,
- kind=artifact.get("kind", "file"),
- description=artifact.get("description", ""),
- )
+ try:
+ original = StageExecutionRecord.from_dict(payload)
+ except (ValueError, TypeError, KeyError, AttributeError):
+ return ResumeDecision(False, "missing original execution evidence; re-running")
+ if original.stage_name != stage_name:
+ return ResumeDecision(False, "original stage name differs; re-running")
+ registered = {a.path for a in original.output_artifacts}
+ if any(not path.exists() for path in registered):
+ return ResumeDecision(False, "a recorded artifact is missing; re-running")
+ if payload.get("artifact_checksums") != artifact_file_checksums(registered):
+ return ResumeDecision(
+ False, "output contents have not been verified or changed; re-running"
)
+ if original.metrics.get("incomplete_reasons") or original.metrics.get("mode") == "fallback":
+ return ResumeDecision(False, "incomplete benchmark; re-running")
- if required_outputs:
- registered = {a.path for a in artifacts}
- if missing_outputs(required_outputs, registered):
- return ResumeDecision(False, "a required stage output is missing or empty; re-running")
- return ResumeDecision(
- True, "resumed from completion marker (fingerprint match)", tuple(artifacts)
- )
+ if missing_outputs(required_outputs, registered):
+ return ResumeDecision(False, "a required stage output is missing or empty; re-running")
+ return ResumeDecision(True, "resumed from completion marker (fingerprint match)", original)
def missing_outputs(required: tuple[Path, ...], registered: set[Path]) -> list[Path]:
diff --git a/src/cellquorum/core/stage.py b/src/cellquorum/core/stage.py
index 5602137a..ff8905e7 100644
--- a/src/cellquorum/core/stage.py
+++ b/src/cellquorum/core/stage.py
@@ -2,7 +2,7 @@
from __future__ import annotations
-from dataclasses import dataclass, field
+from dataclasses import dataclass, field, replace
from datetime import UTC, datetime
from pathlib import Path
from typing import TYPE_CHECKING, Literal, Protocol
@@ -102,7 +102,7 @@ class StageResult:
metrics: JSON-serializable structured metrics for summaries and reports.
"""
- adata: ad.AnnData
+ adata: ad.AnnData | None
artifacts: list[StageArtifact] = field(default_factory=list)
@@ -132,6 +132,14 @@ class StageResult:
checkpoint_path: Path | None = None
+ def require_adata(self) -> ad.AnnData:
+ """Return the output object for a method whose contract requires one."""
+ if self.adata is None:
+ from cellquorum.core.contracts import CellQuorumContractError
+
+ raise CellQuorumContractError("The stage returned no analysis object")
+ return self.adata
+
def __post_init__(self) -> None:
"""
Backfill explicit lifecycle fields from legacy skip metrics.
@@ -426,11 +434,23 @@ class StageExecutionRecord:
def from_dict(cls, row: dict) -> StageExecutionRecord:
"""Restore an execution record written by provenance, including its artifact paths."""
values = dict(row)
+ for key in (
+ "schema_version",
+ "completed_at_utc",
+ "n_output_artifacts",
+ "artifact_checksums",
+ ):
+ values.pop(key, None)
for key in ("started_at_utc", "ended_at_utc"):
values[key] = datetime.fromisoformat(values[key])
for key in ("input_artifacts", "output_artifacts"):
values[key] = [
- StageArtifact(**{**item, "path": Path(item["path"])})
+ StageArtifact(
+ name=item["name"],
+ path=Path(item["path"]),
+ kind=item["kind"],
+ description=item["description"],
+ )
for item in values.get(key, [])
]
if values.get("skip_reason"):
@@ -441,6 +461,14 @@ def from_dict(cls, row: dict) -> StageExecutionRecord:
values["checkpoint_path"] = Path(values["checkpoint_path"])
return cls(**values)
+ def as_reused(self, reason: str) -> StageExecutionRecord:
+ """Keep the original scientific evidence when its outputs are reused."""
+ return replace(
+ self,
+ status="skipped",
+ skip_reason=StageSkipReason(self.stage_name, reason, {"resumed": True}),
+ )
+
@property
def reused_from_checkpoint(self) -> bool:
"""Whether the outputs were reused from a validated completion marker."""
diff --git a/src/cellquorum/core/study_preflight.py b/src/cellquorum/core/study_preflight.py
index 95b65c2f..1a44ca4f 100644
--- a/src/cellquorum/core/study_preflight.py
+++ b/src/cellquorum/core/study_preflight.py
@@ -13,6 +13,7 @@
from cellquorum.config.design import validate_design_against_obs
from cellquorum.config.models import CellQuorumConfig
+from cellquorum.core.anndata_access import obs_frame
from cellquorum.core.context import PipelineContext
from cellquorum.core.exceptions import CellQuorumConfigError
from cellquorum.core.stage import StageArtifact
@@ -97,6 +98,12 @@ def study_blockers(config: CellQuorumConfig, methods: dict[str, list[str]]) -> l
for stage, pkg in [("multicellular_programs", "DIALOGUE"), ("ambient_correction", "SoupX")]:
if stage in methods:
r_packages.add(pkg)
+ if "cytotrace" in methods.get("trajectory", []):
+ from cellquorum.stages.trajectory._cytotrace import worker_issue
+
+ issue = worker_issue()
+ if issue:
+ errors.append(issue)
for method, module in [
("velocity", "scvelo"),
("cellrank", "cellrank"),
@@ -222,7 +229,7 @@ def record_qc_support(
or not config.design.control
):
return None
- obs = adata.obs
+ obs = obs_frame(adata)
keys = [config.design.donor_col, config.design.condition_col]
if any(key not in obs for key in keys):
return None
diff --git a/src/cellquorum/datasets/demo.py b/src/cellquorum/datasets/demo.py
index 1a593a97..ef65a94d 100644
--- a/src/cellquorum/datasets/demo.py
+++ b/src/cellquorum/datasets/demo.py
@@ -15,6 +15,8 @@
import yaml
from scipy import sparse
+from cellquorum.core.anndata_access import copy_x_to_layer
+
N_CELLS = 120
N_GENES = 80
SEED = 71
@@ -60,7 +62,7 @@ def _demo_dataset() -> ad.AnnData:
)
genes = ["MT-CO1", "MT-ND1"] + [f"G{i}" for i in range(N_GENES - 2)]
data = ad.AnnData(sparse.csr_matrix(counts), obs=obs, var=pd.DataFrame(index=genes))
- data.layers["counts"] = data.X.copy()
+ copy_x_to_layer(data, "counts")
return data
diff --git a/src/cellquorum/io/anndata.py b/src/cellquorum/io/anndata.py
index fd4a497c..3d4c9c2b 100644
--- a/src/cellquorum/io/anndata.py
+++ b/src/cellquorum/io/anndata.py
@@ -11,6 +11,7 @@
import numpy as np
import pandas as pd
+from cellquorum.core.anndata_access import obs_frame
from cellquorum.core.contracts.layer_tags import get_layer_tag, set_layer_tag
from cellquorum.core.exceptions import CellQuorumDataError
@@ -255,7 +256,7 @@ def _load_adata_subset(
mask = np.ones(n_before, dtype=bool)
else:
column_as_str = _require_obs_column(
- backed.obs, subset_column, setting="input.subset.column"
+ obs_frame(backed), subset_column, setting="input.subset.column"
)
wanted = {str(value) for value in subset_values}
mask = column_as_str.isin(wanted).to_numpy()
@@ -266,7 +267,7 @@ def _load_adata_subset(
if exclude_column is not None:
exclude_as_str = _require_obs_column(
- backed.obs, exclude_column, setting="input.exclude.column"
+ obs_frame(backed), exclude_column, setting="input.exclude.column"
)
unwanted = {str(value) for value in (exclude_values or [])}
@@ -295,7 +296,7 @@ def _load_adata_subset(
if agreement_column is not None:
assert column_as_str is not None # guarded by load_adata
agreement_as_str = _require_obs_column(
- backed.obs, agreement_column, setting="input.subset.require_agreement"
+ obs_frame(backed), agreement_column, setting="input.subset.require_agreement"
)
# Guard the VOCABULARY before the cells. A label the second annotation
@@ -381,7 +382,7 @@ def declare_layer_kinds(adata: ad.AnnData, kinds: Mapping[str, str]) -> None:
if layer not in adata.layers:
raise AnnDataLoadError(
f"input.layer_kinds names layer '{layer}', which is not in the file "
- f"(layers: {sorted(k for k in adata.layers if k is not None)})"
+ f"(layers: {sorted(k for k in adata.layers.keys() if k is not None)})"
)
existing = get_layer_tag(adata, layer)
if existing is not None and existing["kind"] != kind:
diff --git a/src/cellquorum/io/study_inputs.py b/src/cellquorum/io/study_inputs.py
index 771f9db9..533172cd 100644
--- a/src/cellquorum/io/study_inputs.py
+++ b/src/cellquorum/io/study_inputs.py
@@ -9,6 +9,8 @@
from scipy.io import mmread
from scipy.sparse import csr_matrix
+from cellquorum.core.anndata_access import copy_x_to_layer, memory_matrix
+
def _first(directory: Path, names: tuple[str, ...]) -> Path:
for name in names:
@@ -123,9 +125,10 @@ def import_libraries(source: Path, destination: Path) -> pd.DataFrame:
)
if "cellquorum_input_X" in data.layers:
raise ValueError(f"Reserved layer cellquorum_input_X already exists in {sample}.")
- data.layers["cellquorum_input_X"] = data.X.copy()
+ copy_x_to_layer(data, "cellquorum_input_X")
data.X = data.layers[row["counts_layer"]].copy()
- values = data.X.data if hasattr(data.X, "tocsr") else np.asarray(data.X)
+ matrix = memory_matrix(data.X)
+ values = matrix if isinstance(matrix, np.ndarray) else matrix.data
if not (np.isfinite(values) & (values >= 0) & (values == np.floor(values))).all():
raise ValueError(
f"{sample} needs nonnegative integer raw counts in X or a counts_layer."
@@ -137,7 +140,8 @@ def import_libraries(source: Path, destination: Path) -> pd.DataFrame:
if not metadata.index.is_unique or set(metadata.index) != set(data.obs_names):
raise ValueError(f"Cell metadata barcodes must match {sample} exactly.")
aligned = metadata.loc[data.obs_names]
- for key in aligned:
+ for column in aligned.columns:
+ key = str(column)
if key in data.obs and not data.obs[key].astype(str).equals(
aligned[key].astype(str)
):
diff --git a/src/cellquorum/methods/context_access.py b/src/cellquorum/methods/context_access.py
index d889ad72..f4bfd7b6 100644
--- a/src/cellquorum/methods/context_access.py
+++ b/src/cellquorum/methods/context_access.py
@@ -1,11 +1,4 @@
-"""Resolve a stage's config sub-block from a pipeline context.
-
-The real ``PipelineContext.config`` is a ``CellQuorumConfig`` pydantic object, so
-a stage reads its own settings via ``getattr(config, stage_name).model_dump()``.
-Unit tests, however, often pass a plain dict context. This helper accepts either
-and always returns a plain dict, so ``MethodDispatchStage`` subclasses work in
-both settings without branching.
-"""
+"""Resolve study and stage settings for analysis methods."""
from __future__ import annotations
@@ -49,4 +42,27 @@ def resolve_stage_config(context: PipelineContext, stage_name: str) -> dict:
return {}
-__all__ = ["resolve_stage_config"]
+def with_study_contrast(context: PipelineContext, stage_config: dict) -> dict:
+ """Fill missing comparison settings from the cohort and primary study design."""
+ from cellquorum.config.cohort import resolve_design_columns
+
+ settings = dict(stage_config)
+ config = getattr(context, "config", None)
+ design = getattr(config, "design", None)
+ if design is None:
+ return settings
+
+ donor, condition = resolve_design_columns(config, donor_key=None, condition_key=None)
+ for key, value in (
+ ("donor_col", donor),
+ ("condition_col", condition),
+ ("case", getattr(design, "case", None)),
+ ("control", getattr(design, "control", None)),
+ ):
+ if not settings.get(key):
+ settings[key] = value
+ settings.setdefault("paired", bool(getattr(design, "paired", False)))
+ return settings
+
+
+__all__ = ["resolve_stage_config", "with_study_contrast"]
diff --git a/src/cellquorum/methods/stage_base.py b/src/cellquorum/methods/stage_base.py
index 15a9f2d9..a02bc3a7 100644
--- a/src/cellquorum/methods/stage_base.py
+++ b/src/cellquorum/methods/stage_base.py
@@ -17,6 +17,8 @@ class that satisfies the existing ``PipelineStage`` Protocol.
from cellquorum.methods.registry import METHOD_REGISTRY, MethodRegistry
if TYPE_CHECKING:
+ import anndata as ad
+
from cellquorum.config.models import CellQuorumConfig
from cellquorum.core.context import PipelineContext
@@ -124,7 +126,7 @@ def run(self, context: PipelineContext) -> StageResult:
def _run_single_method(
self,
- adata: object,
+ adata: ad.AnnData,
method_config: dict,
context: PipelineContext,
) -> StageResult | MethodSkip:
@@ -152,7 +154,7 @@ def _run_single_method(
def _run_methods_list(
self,
- adata: object,
+ adata: ad.AnnData,
stage_config: dict,
methods: list[dict],
context: PipelineContext,
@@ -206,7 +208,7 @@ def _run_methods_list(
# Validate and thread this method's AnnData into the next method.
self._validate_output(outcome)
- current = outcome.adata
+ current = outcome.require_adata()
notes.extend(outcome.notes)
warnings.extend(outcome.warnings)
artifacts.extend(outcome.artifacts)
@@ -281,6 +283,18 @@ def _validate_output(self, result: StageResult) -> None: # noqa: B027
"""
+class DefaultMethodStage(MethodDispatchStage):
+ """Dispatch a named method, using the stage's declared default when omitted."""
+
+ @property
+ @abstractmethod
+ def default_method(self) -> str:
+ """Registry name used by a study that omits the method key."""
+
+ def _select_method_name(self, config: dict) -> str:
+ return config.get("method", self.default_method)
+
+
# Cohort attributes that overlay onto same-named stage config keys. These are
# the structural obs columns a dataset declares once (config.cohort). When a
# cohort field is set it takes precedence, so the biological structure is
@@ -329,4 +343,4 @@ def _apply_cohort_overlay(context: PipelineContext, stage_config: dict) -> dict:
return overlay_cohort_keys(getattr(context, "config", None), stage_config)
-__all__ = ["MethodDispatchStage"]
+__all__ = ["DefaultMethodStage", "MethodDispatchStage"]
diff --git a/src/cellquorum/stages/ambient_correction/soupx.py b/src/cellquorum/stages/ambient_correction/soupx.py
index 5c2ed562..6559be19 100644
--- a/src/cellquorum/stages/ambient_correction/soupx.py
+++ b/src/cellquorum/stages/ambient_correction/soupx.py
@@ -16,6 +16,7 @@
import scipy.sparse as sp
from cellquorum.backends.script_paths import r_script_path
+from cellquorum.core.anndata_access import copy_x_to_layer
from cellquorum.core.exceptions import CellQuorumBackendError, CellQuorumDataError
# Path to the bundled SoupX R script.
@@ -173,7 +174,7 @@ def import_corrected_matrix(out_dir: str | Path, sample_id: str) -> ad.AnnData:
# Build the AnnData; namespace barcodes by sample.
adata = ad.AnnData(X=sp.csr_matrix(mat))
- adata.var_names = genes
+ adata.var_names = genes.tolist()
if (d / "gene_ids.tsv").is_file():
adata.var["gene_ids"] = pd.read_csv(d / "gene_ids.tsv", header=None, dtype=str)[
0
@@ -184,7 +185,7 @@ def import_corrected_matrix(out_dir: str | Path, sample_id: str) -> ad.AnnData:
# Record sample_id as an explicit obs column (robust — do NOT reconstruct it
# downstream by splitting namespaced barcodes, which can themselves contain '_').
adata.obs["sample_id"] = sample_id
- adata.layers["counts"] = adata.X.copy()
+ copy_x_to_layer(adata, "counts")
return adata
diff --git a/src/cellquorum/stages/ambient_correction/stage.py b/src/cellquorum/stages/ambient_correction/stage.py
index 65e53ce4..6463123f 100644
--- a/src/cellquorum/stages/ambient_correction/stage.py
+++ b/src/cellquorum/stages/ambient_correction/stage.py
@@ -19,6 +19,7 @@
import anndata as ad
import pandas as pd
+from cellquorum.core.anndata_access import memory_matrix
from cellquorum.core.contracts import CellQuorumContractError
from cellquorum.core.fingerprint import resource_checksum
from cellquorum.core.stage import StageArtifact, StageResult
@@ -187,12 +188,12 @@ def run(self, context: PipelineContext) -> StageResult:
raise CellQuorumContractError(
"Rerun SoupX with resume: false to record stable gene IDs."
)
- library.var_names = library.var["gene_ids"].astype(str)
+ library.var_names = library.var["gene_ids"].astype(str).tolist()
library.obs_names = (
library.obs["sample_id"].astype(str)
+ ":"
+ library.obs["barcode"].astype(str)
- )
+ ).tolist()
combined = ad.concat(corrected_libraries, join="inner")
ids = adata.var["gene_ids"].astype(str)
if set(combined.obs_names) != set(adata.obs_names) or set(
@@ -207,7 +208,7 @@ def run(self, context: PipelineContext) -> StageResult:
"Choose a fresh input. The uncorrected_counts layer already exists."
)
corrected_adata.layers["uncorrected_counts"] = corrected_adata.X.copy()
- corrected_adata.X = combined[adata.obs_names, ids].X.copy()
+ corrected_adata.X = memory_matrix(combined[adata.obs_names, ids].X).copy()
corrected_adata.layers["counts"] = corrected_adata.X.copy()
else:
corrected_adata = ad.concat(corrected_libraries, join="outer")
diff --git a/src/cellquorum/stages/annotation/adjudication/evidence.py b/src/cellquorum/stages/annotation/adjudication/evidence.py
index 007945d5..4f4c18ab 100644
--- a/src/cellquorum/stages/annotation/adjudication/evidence.py
+++ b/src/cellquorum/stages/annotation/adjudication/evidence.py
@@ -5,6 +5,7 @@
import anndata as ad
import pandas as pd
+from cellquorum.core.anndata_access import obs_frame
from cellquorum.stages.annotation.adjudication.adjudicate import ClusterEvidence
from cellquorum.stages.annotation.adjudication.config import AdjudicationConfig
@@ -39,7 +40,7 @@ def build_cluster_evidence_table(
raise KeyError(f"Missing required adjudication obs column(s): {missing}.")
# Work with strings for stable grouping/reporting.
- obs = adata.obs.copy()
+ obs = obs_frame(adata).copy()
obs[config.cluster_key] = obs[config.cluster_key].astype(str)
obs[donor_key] = obs[donor_key].astype(str)
obs[condition_key] = obs[condition_key].astype(str)
diff --git a/src/cellquorum/stages/annotation/adjudication/stage.py b/src/cellquorum/stages/annotation/adjudication/stage.py
index 37004c68..38cbf038 100644
--- a/src/cellquorum/stages/annotation/adjudication/stage.py
+++ b/src/cellquorum/stages/annotation/adjudication/stage.py
@@ -8,6 +8,7 @@
import pandas as pd
+from cellquorum.config.cohort import resolve_design_columns
from cellquorum.core.artifacts import ArtifactManager
from cellquorum.core.stage import StageResult
from cellquorum.core.stage_catalog import register_stage
@@ -235,38 +236,12 @@ def resolve_design_keys(
if config.donor_key and config.condition_key:
return config.donor_key, config.condition_key
-
- context_config = getattr(context, "config", None)
-
- # Cohort schema wins when it declares the structural keys (declare-once),
- # matching population_identity's resolve_design_keys -- otherwise a real
- # cohort whose donor column is e.g. "donor_id" silently falls through to
- # the "patient_id" default below instead of the column that actually exists.
- from cellquorum.config.cohort import resolve_cohort_key
-
- donor_key = resolve_cohort_key(context_config, attr="donor_key", stage_value=config.donor_key)
- condition_key = resolve_cohort_key(
- context_config, attr="condition_key", stage_value=config.condition_key
+ return resolve_design_columns(
+ getattr(context, "config", None),
+ donor_key=config.donor_key,
+ condition_key=config.condition_key,
)
- design = getattr(context_config, "design", None)
- if isinstance(context_config, Mapping):
- design = context_config.get("design", {})
-
- if donor_key is None:
- donor_key = (
- design.get("donor_col", "patient_id")
- if isinstance(design, Mapping)
- else getattr(design, "donor_col", "patient_id")
- )
- if condition_key is None:
- condition_key = (
- design.get("condition_col", "condition")
- if isinstance(design, Mapping)
- else getattr(design, "condition_col", "condition")
- )
- return donor_key, condition_key
-
__all__ = [
"AdjudicationStage",
diff --git a/src/cellquorum/stages/annotation/celltypist_method.py b/src/cellquorum/stages/annotation/celltypist_method.py
index 526a5f42..9354ceca 100644
--- a/src/cellquorum/stages/annotation/celltypist_method.py
+++ b/src/cellquorum/stages/annotation/celltypist_method.py
@@ -81,9 +81,9 @@ def _run(
X=adata.layers[counts_layer].copy(),
obs=pd.DataFrame(index=adata.obs_names),
var=pd.DataFrame(index=adata.var_names),
- obsp={key: adata.obsp[key] for key in adata.obsp},
uns={"neighbors": adata.uns["neighbors"]} if "neighbors" in adata.uns else {},
)
+ work.obsp.update(adata.obsp.items())
sc.pp.normalize_total(work, target_sum=1e4)
sc.pp.log1p(work)
diff --git a/src/cellquorum/stages/annotation/diagnostics/scdiagnostics_method.py b/src/cellquorum/stages/annotation/diagnostics/scdiagnostics_method.py
index 26e14633..38a68354 100644
--- a/src/cellquorum/stages/annotation/diagnostics/scdiagnostics_method.py
+++ b/src/cellquorum/stages/annotation/diagnostics/scdiagnostics_method.py
@@ -12,6 +12,7 @@
import pandas as pd
from cellquorum.backends.script_paths import r_script_path
+from cellquorum.core.anndata_access import obs_frame
from cellquorum.core.contracts import DataContract
from cellquorum.core.exceptions import CellQuorumBackendError
from cellquorum.core.h5ad_io import write_h5ad
@@ -334,7 +335,10 @@ def _run(
description="Per-cell scDiagnostics confidence metrics.",
),
*_write_figures(
- obs=result_adata.obs, cell_type_col=cell_type_col, context=context, config=config
+ obs=obs_frame(result_adata),
+ cell_type_col=cell_type_col,
+ context=context,
+ config=config,
),
]
@@ -538,7 +542,7 @@ def _run_probability_entropy_only(
)
figure_artifacts = _write_figures(
- obs=result_adata.obs, cell_type_col=cell_type_col, context=context, config=config
+ obs=obs_frame(result_adata), cell_type_col=cell_type_col, context=context, config=config
)
return StageResult(
@@ -691,8 +695,8 @@ def _write_query_h5ad(
adata = adata[:, shared]
query = ad.AnnData(X=adata.layers[expression_layer].copy())
- query.obs_names = adata.obs_names
- query.var_names = adata.var_names
+ query.obs_names = adata.obs_names.tolist()
+ query.var_names = adata.var_names.tolist()
# Written as plain strings. A pandas Categorical becomes an h5ad categorical, which
# zellkonverter reads as an R factor, which scDiagnostics' `projectPCA` reports as the
# factor's integer CODES — so its `cell_type == "