Version 5.7.2

New features

(none!)

Bug-fixes and performance

• fix: make Library.eq safe for invalid comparisons by @Rachit931 in #1565
• fix: respect entry bounds for RNTuple.iterate and add of regressions tests for out of bounds by @KaranSinghDev in #1573
• fix: resolve issues with pyroot buffer reallocation by @ariostas in #1569
• fix: GPU interpretation of RNTuple data working and implement nvCOMP decompression with nvCOMP library by @fstrug in #1557
• fix: allow mkrntuple to accept wider range of objects that can be reduced to RecordArray/RecordForm by @ariostas in #1577
• fix: ensure RNTuple _from_zigzag applies bit shift on unsigned integers by @wdconinc in #1593
• fix: make inherited fields accessible by @ariostas in #1589
• fix: reading RNTuple fields into NumPy by @ariostas in #1591
• refactor: Removes optional handling for required dependencies. by @Rachit931 in #1579

Other

• chore(deps): bump actions/attest-build-provenance from 3.1.0 to 3.2.0 in the actions group by @dependabot[bot] in #1566
• chore: update pre-commit hooks by @pre-commit-ci[bot] in #1567
• docs: add AI-assisted contributions guidelines by @ianna in #1562
• chore: skip broken fsspec version by @ariostas in #1583
• chore: support RNTuple v1.0.1.0 format (attribute set record frame) by @wdconinc in #1580
• chore(deps): bump the actions group with 2 updates by @dependabot[bot] in #1587
• chore: update pre-commit hooks by @pre-commit-ci[bot] in #1588
• chore: add codecov by @ariostas in #1590
• docs: add wdconinc as a contributor for code by @allcontributors[bot] in #1584
• docs: add KaranSinghDev as a contributor for code by @allcontributors[bot] in #1585
• docs: RNTuple GPU reading by @fstrug in #1595

New Contributors

• @wdconinc made their first contribution in #1580
• @KaranSinghDev made their first contribution in #1573

Full Changelog: v5.7.1...v5.7.2