LDaCA Wordflow

Wordflow is a text-analysis application built with Rust, DuckDB, React and Tauri. Each desktop project window owns an embedded Axum backend and a .wfpj DuckDB file.

Packages

• backend/: native project lifecycle, SQL, graph metadata and imports.
• ldaca-rs/: reusable text algorithms, model inference and ONI/RO-Crate data access.
• frontend/: React interface and Tauri desktop host.
• polars-text: independent Polars/Python adapter over ldaca-rs.
• ldaca-analytics-sample-data/: remote sample data.
• ldaca-wordflow-docs/: published user-documentation mirror.
• archive/: retired Python backend and associated infrastructure, preserved for reference rather than execution.

Development

pnpm install
pnpm dev                 # Rust backend and browser preview
pnpm dev:desktop         # Tauri desktop application
pnpm build:desktop:mac   # Regular macOS application bundle

Rust/Cargo is required. The active application does not need a Python environment or a Data Root. The browser preview uses the native project interface; the old FastAPI server workflow is archived. The optional polars-text adapter is developed in its own repository and resolves ldaca-rs from a locked Git dependency.

See the engineering index, development instructions, and desktop runtime instructions.