-
Notifications
You must be signed in to change notification settings - Fork 945
Forward-merge branch-25.06 into branch-25.08 #18713
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
Support for Python 3.13 was added in #18648, however the `cudf-pandas-unit-test` job requires some extra attention to refactor how it pulls in compatible versions of Pandas (Python 3.13 support was added to Pandas in version 2.2.3). Until someone has time to investigate and fix that, moving this job away from the Python 3.13 workflows to get nightlies passing. Authors: - Gil Forsyth (https://github.com/gforsyth) - David Wendt (https://github.com/davidwendt) - Bradley Dice (https://github.com/bdice) Approvers: - Bradley Dice (https://github.com/bdice) - GALI PREM SAGAR (https://github.com/galipremsagar) - Vyas Ramasubramani (https://github.com/vyasr) URL: #18692
FAILURE - Unable to forward-merge due to an error, manual merge is necessary. Do not use the IMPORTANT: When merging this PR, do not use the auto-merger (i.e. the |
Currently we print a notice when prefetching is enabled and we try to use a type that cannot be safely prefetched (essentially any nested type). We should not be printing unconditionally. Logging is more appropriate. Authors: - Vyas Ramasubramani (https://github.com/vyasr) - GALI PREM SAGAR (https://github.com/galipremsagar) Approvers: - GALI PREM SAGAR (https://github.com/galipremsagar) - Bradley Dice (https://github.com/bdice) - Vukasin Milovanovic (https://github.com/vuule) URL: #18711
Resolves: #18569 This PR enables strict pass/fail for `pandas-tests` job by skipping all know failures. Authors: - GALI PREM SAGAR (https://github.com/galipremsagar) - Bradley Dice (https://github.com/bdice) Approvers: - Bradley Dice (https://github.com/bdice) URL: #18641
Add tests that are parameterized on the compression codec and the implementation (nvCOMP, internal kernel, host) to ensure all available implementations are tested for all supported codecs. More tests will be added as issues with non-nvCOMP implementations are fixed. Also: tests uncovered another bug in the Parquet writer; fixed it in this PR. Authors: - Vukasin Milovanovic (https://github.com/vuule) Approvers: - David Wendt (https://github.com/davidwendt) - MithunR (https://github.com/mythrocks) URL: #18690
- Closes #17123 Authors: - Lawrence Mitchell (https://github.com/wence-) Approvers: - Vyas Ramasubramani (https://github.com/vyasr) URL: #18634
This PR fully resolves #18665 by moving the test to benchmarks and only verifying the correctness result on a smaller subset of values in the regular pytest suites. Authors: - GALI PREM SAGAR (https://github.com/galipremsagar) Approvers: - Bradley Dice (https://github.com/bdice) URL: #18715
Follow on to #18692 to include same change for nightly tests. Authors: - David Wendt (https://github.com/davidwendt) Approvers: - Bradley Dice (https://github.com/bdice) - GALI PREM SAGAR (https://github.com/galipremsagar) URL: #18717
Part of the strings split refactor work. This change reworks the `split_helper` internal utility for better reuse with split-on-whitespace and tokenizers (follow on work). Authors: - David Wendt (https://github.com/davidwendt) Approvers: - Vyas Ramasubramani (https://github.com/vyasr) - Devavret Makkar (https://github.com/devavret) - Nghia Truong (https://github.com/ttnghia) - Shruti Shivakumar (https://github.com/shrshi) URL: #18575
This PR does two things: - Replaces the `parquet_bocksize` configuration options with a general `target_partition_size` configuration that can be used by other IO functions (and possibly for reduction/rebalancing/rechunking in the future). - Updates the `rapidsmpf` spilling wrapper to ask for headroom before executing tasks. In local experiments, I found that this made some fragile PDS-H queries much more stable. **TODO** (Before I remove the PR from "draft") - [x] Align with #18623 - [x] Re-run local PDS-H experiments at sf1k and sf3k with this branch to double-check behavior. Authors: - Richard (Rick) Zamora (https://github.com/rjzamora) Approvers: - Tom Augspurger (https://github.com/TomAugspurger) - Peter Andreas Entschev (https://github.com/pentschev) - Mads R. B. Kristensen (https://github.com/madsbk) URL: #18669
Teeing up this "fix" for the proposed change in rapidsai/rapidsmpf#256 Once that PR is merged, we will want to get this in asap to keep `rapidsmpf` shuffling from breaking. We can update `Sort` in a follow-up PR. Authors: - Richard (Rick) Zamora (https://github.com/rjzamora) Approvers: - Tom Augspurger (https://github.com/TomAugspurger) URL: #18720
Forward-merge triggered by push to branch-25.06 that creates a PR to keep branch-25.08 up-to-date. If this PR is unable to be immediately merged due to conflicts, it will remain open for the team to manually merge. See forward-merger docs for more info.