All pull requests

Pull requests

• Replace BioStructure/BioSequence with a dependency-free Structure feature

  #8895·kierandidi opened Oct 25, 2026·

  ·

• docs: propose evidence-first evolution intake and licensing gates

  #8894·fisallllll280-code opened Oct 25, 2026·

  ·

• Fix xgetsize fallback to count bytes instead of decoded characters

  #8893·Sonica-B opened Oct 25, 2026·

  ·

• Forward ignore_decode_attribute in nested require_decoding calls

  #8892·udsy19 opened Oct 24, 2026·

  ·

• Don't infer the Harbor builder from unrelated .md and .toml files

  #8891·NgoQuocViet2001 opened Oct 24, 2026·

  ·

• keep nullable integers intact in to_pandas

  #8890·chessing234 opened Oct 24, 2026·

  ·

• Keep row validity and payloads in fixed-shape ArrayXD pandas round trips

  #8889·choksi2212 opened Oct 24, 2026·

  ·
  1

  2

• Preserve selected streaming rows across concatenated epochs

  #8887·sjdoane opened Oct 23, 2026·

  ·

• Fix SQLAlchemy query cache collisions for bound parameters

  #8886·gorkadabo1 opened Oct 23, 2026Contributor·

  ·
  1

  1

• Close Parquet export writers with a context manager

  #8885·wpyaoxueros opened Oct 23, 2026·

  ·
  1

• Reject duplicate dataset card configuration names

  #8883·JackyHsu-X opened Oct 23, 2026·

  ·

• include bound SQL values in the dataset cache key

  #8881·chessing234 opened Oct 23, 2026·

  ·
  1

  3

• Fix save_to_disk sending the whole table to every worker for in-memory datasets

  #8880·hzmsh opened Oct 23, 2026Contributor·

  ·
  1

• Fix pickle RCE in pandas dataset builder (CVE-2026-57845)

  #8879·SaiTeja-Erukude opened Oct 23, 2026·

  ·

• Fix decoding a streamed PDF dataset twice after shuffle, take or filter

  #8878·cwklurks opened Oct 22, 2026Contributor·5 tasks done·

  ·

• Faster IterableDataset iteration

  #8876·vsaraikin opened Oct 22, 2026·

  ·
  1

• Honor table policies when exporting empty datasets to SQL

  #8875·emme1t opened Oct 22, 2026·

  ·

• Preserve unrestricted filters when extending data-file patterns

  #8874·nawaaaaaAaar opened Oct 22, 2026·

  ·
  1

• Fix doubled words in load.py docstring and process docs

  #8873·aarushkandukoori opened Oct 22, 2026Contributor·

  ·

• Preserve grouped streaming batches when resuming checkpoints

  #8872·liangqianxing opened Oct 22, 2026·

  ·

• Honor batch_size when exporting an IterableDataset to Parquet

  #8871·liangqianxing opened Oct 22, 2026·

  ·

• Avoid copying WebDataset binary payloads during schema inference

  #8870·cola1917 opened Oct 22, 2026·

  ·
  1

• Raise on duplicate column names in Dataset.select_columns

  #8868·astroknot-sheep opened Oct 21, 2026·

  ·

• Preserve rows when concatenating pandas ArrayXD batches

  #8867·pavanmudumba opened Oct 20, 2026Contributor·

  ·
  1

• Fix forwarding custom User-Agent to fsspec HTTP requests

  #8866·MonikaDvorackova opened Oct 20, 2026·

  ·