Skip to content

Name a from_generator cache by a given fingerprint alone, as in Python - #112

Merged
CarloLucibello merged 1 commit into
mainfrom
cl/share-fingerprint-cache
Oct 20, 2026
Merged

CarloLucibello merged 1 commit into
mainfrom
cl/share-fingerprint-cache

Conversation

@CarloLucibello

Copy link
Copy Markdown
Collaborator

Dataset(stream; fingerprint, cache_dir) / Dataset.from_generator(...; fingerprint) hashed the fingerprint together with features and split (generator/default-<hash>/), while Python's from_generator names the cache by the fingerprint alone (generator/default-fingerprint=<fp>/). So Julia and Python code with the same fingerprint built separate, identical caches.

A given fingerprint now goes to datasets' own keyword, so both share one cache (as long as the split matches, as in Python). The content-hash path for Julia sources is unchanged. Datasets cached under a fingerprint by 0.4.4 are rebuilt once.

Also dates the [0.4.4] changelog section, which the release commit missed, and opens [0.4.5].

🤖 Generated with Claude Code

Forward a user-given `fingerprint` to `datasets`' own keyword instead of
hashing it with `features` and `split`, so Julia and Python code calling
`from_generator` with the same `fingerprint`, `cache_dir` and `split`
share one cache. Also date the 0.4.4 changelog section.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@CarloLucibello
CarloLucibello merged commit 593eff9 into main Oct 20, 2026
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant