jzombie / rust-triplets / 29603096773

96%

Build Type

push

github

Committed by web-flow

Commit Message

v0.25.0-alpha (#132)

* Bump parquet from 58.3.0 to 59.1.0 (#126)

Bumps [parquet](https://github.com/apache/arrow-rs) from 58.3.0 to 59.1.0.
- [Release notes](https://github.com/apache/arrow-rs/releases)
- [Changelog](https://github.com/apache/arrow-rs/blob/main/CHANGELOG.md)
- [Commits](https://github.com/apache/arrow-rs/compare/58.3.0...59.1.0)

---
updated-dependencies:
- dependency-name: parquet
  dependency-version: 59.1.0
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

* Add `Sampler` mixed strategy tests (#127)

* Remove `hf-hub` dependency and legacy `datasets-server` fallback paths  (#130)

Why: Full control over download and cache handling.

Also, fixes new 503 errors which were starting to originate from deprecated
`datasets-server` fallbacks.

Replace the datasets-server /parquet, /size, and /info endpoints with
the Hub API tree endpoint (/api/datasets/{dataset}/tree/main). Remove
the hf-hub crate dependency entirely, eliminating the three legacy
fallback paths: sibling-based candidate resolution, ClassLabel /info
resolution, and global row count /size queries.
Key changes:
- remote_url_for_candidate handles bare relative paths, full HTTP URLs,
  and url::-prefixed candidates
- candidate_target_path resolves CDN /resolve/ URLs and bare relative
  paths
- Tree endpoint scoped to config subdirectory; falls back to root for
  "default"/empty config
- Path traversal guard rejects ParentDir/Prefix components
- HEAD fallback in download_and_materialize_shard_with_runtime uses
  fetch_remote_size_with_runtime
- target_matches_expected_size rejects 0-byte files when expected_bytes
  is None
- Mock servers return full CDN URLs; HfMockServer uses org/dataset
  dataset name
- Removed: label_maps, size_endpoint, info_endpoint, known_total_rows(),
  collect_candidates_from_sibling... (continued)

Coverage Stats

5137 of 5662 new or added lines in 35 files covered. (90.73%)

11 existing lines in 4 files now uncovered.

21068 of 21987 relevant lines covered (95.82%)

5318.83 hits per line