You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Bumps the version to 0.7.0, the first release on DataFusion 55. It's the release xqlsystems/datafusion-accelerated#41 needs, so the nanoGPT step can run against released xarray-sql. Closes#310.
Changes: the version in Cargo.toml and Cargo.lock, and the wheel name in docs/browser.md's local-testing example.
Checked before release: the ML demos (tests/test_ml_demos.py: 01_attention, and 02_nanogpt with Muon and AdamW) pass on main, on Linux x86_64 (i7-8700) with Python 3.12.
The locked versions: ddxdb 0.4.6 and einfold 0.1.0rc2.
The newest releases: ddxdb 0.4.9 and einfold 0.2.0, which pip install "xarray-sql[ml]" resolves to today. Both with datafusion 55.0.0, duckdb 1.5.6 and jax 0.12.0.1.
Release notes (draft, for the GitHub release):
xarray-sql 0.7.0 moves to DataFusion 55. It adds gradients in SQL with the new ml extra, and prepared queries that are planned once and run many times. xarray-sql, ddxdb (ddx's Python package) and einfold can now be installed together, on one version of DataFusion.
New
Gradients in SQL: pip install "xarray-sql[ml]". It installs ddx, which differentiates SQL, so a query can call grad(), vjp() and jvp() over xarray Datasets and any DataFusion table. Two demos in benchmarks/ml train models this way and check their results against JAX: attention, and a nanoGPT training step (Add an ml extra for gradients in SQL, and two ML demos #269).
Preparing a query that takes a gradient now runs it once, to find the tables it reads. Writing new data to a table the query doesn't read raises an error, as it already did for queries without gradients (Reuse gradient plans and reduce Dataset registration overhead #301).
This release bump updates the version and the local demo wheel name in the docs. It has no meaningful behavior risk, so it is ready to merge.
Pre-merge checks: 3 · 1
Failed checks: 1 inconclusive
Check name
Status
Explanation
Resolution
Linked Issues check
Issue #310 asks for a release on DataFusion 55 that lets xarray-sql, ddxdb and einfold install together. It states that main pins einfold<0.2, and that allowing einfold 0.2.0 is the needed outcome. …
Confirm the einfold constraint in the Python package metadata (for example pyproject.toml, [ml] extra) at the reviewed head 95c1d34. The constraint must allow einfold 0.2.0 and must not pin einfold<0.2. Also confirm that the DataFusio…
Passed checks: 3 passed
Check name
Status
Explanation
Description Check
Check skipped - CodeRabbit’s high-level summary is enabled.
Title check
The title "Release 0.7.0" names the version bump, which is the main change in the changeset (Cargo.toml and docs/browser.md wheel filename both move from 0.6.0 to 0.7.0). It is short and specific, but…
Out of Scope Changes check
The reviewed changes are a version bump in Cargo.toml from 0.6.0 to 0.7.0 and a wheel filename update in docs/browser.md from 0.6.0 to 0.7.0. Both changes match the stated purpose of the PR, which is …
Full details: Title check
Explanation
The title "Release 0.7.0" names the version bump, which is the main change in the changeset (Cargo.toml and docs/browser.md wheel filename both move from 0.6.0 to 0.7.0). It is short and specific, but it does not say what the release contains (DataFusion 55 support), so it is brief rather than fully descriptive. It is related to the main change, so the check passes.
Full details: Out of Scope Changes check
Explanation
The reviewed changes are a version bump in Cargo.toml from 0.6.0 to 0.7.0 and a wheel filename update in docs/browser.md from 0.6.0 to 0.7.0. Both changes match the stated purpose of the PR, which is the 0.7.0 release. The release notes in the PR description describe features (ml extra, prepare, threads, performance) that the summary does not show as changed here. These features appear to come from earlier merged work, not from this diff, so they are not out-of-scope changes in this PR.
Full details: Linked Issues check
Explanation
Issue #310 asks for a release on DataFusion 55 that lets xarray-sql, ddxdb and einfold install together. It states that main pins einfold<0.2, and that allowing einfold 0.2.0 is the needed outcome. The PR changes only the version in Cargo.toml (0.6.0 to 0.7.0) and one wheel name in docs/browser.md. The summary shows no change to the Python dependency constraints. The PR description says pip install "xarray-sql[ml]" already resolves einfold 0.2.0. The summary does not show whether the einfold<0.2 pin is absent at the reviewed head. Cargo.lock is ignored, so the locked versions cannot be checked from the excluded file. The version bump and release notes address the release part of #310. The einfold constraint part cannot be confirmed from the available evidence.
Resolution
Confirm the einfold constraint in the Python package metadata (for example pyproject.toml, [ml] extra) at the reviewed head 95c1d34. The constraint must allow einfold 0.2.0 and must not pin einfold<0.2. Also confirm that the DataFusion pin is 55.x in the same metadata. Cargo.lock is excluded by the !**/*.lock rule, so it cannot serve as evidence here.
Finishing Touches
Generate unit tests (beta)
Commit to this branch
Create a new PR
Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.
File: No configuration file from this repository applied
Commit: Unavailable
Layers: defaults
Review profile: CHILL
Plan: Advanced
Run ID: 9da6607c-dbe1-4d26-95d4-dc83877aaf94
Commits
Reviewing files that changed from the base of the PR and between 679cbda and 95c1d34.
Files ignored due to path filters: 1
Cargo.lock is excluded by !**/*.lock
Files selected for processing: 2
Cargo.toml
docs/browser.md
Included review availability: This review used your included allowance. Your plan provides up to 2 included reviews per hour; 1 remain after this review.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Bumps the version to 0.7.0, the first release on DataFusion 55. It's the release xqlsystems/datafusion-accelerated#41 needs, so the nanoGPT step can run against released xarray-sql. Closes #310.
Changes: the version in
Cargo.tomlandCargo.lock, and the wheel name indocs/browser.md's local-testing example.Checked before release: the ML demos (
tests/test_ml_demos.py: 01_attention, and 02_nanogpt with Muon and AdamW) pass on main, on Linux x86_64 (i7-8700) with Python 3.12.pip install "xarray-sql[ml]"resolves to today. Both with datafusion 55.0.0, duckdb 1.5.6 and jax 0.12.0.1.Release notes (draft, for the GitHub release):
xarray-sql 0.7.0 moves to DataFusion 55. It adds gradients in SQL with the new
mlextra, and prepared queries that are planned once and run many times. xarray-sql, ddxdb (ddx's Python package) and einfold can now be installed together, on one version of DataFusion.New
pip install "xarray-sql[ml]". It installs ddx, which differentiates SQL, so a query can callgrad(),vjp()andjvp()over xarray Datasets and any DataFusion table. Two demos inbenchmarks/mltrain models this way and check their results against JAX: attention, and a nanoGPT training step (Add an ml extra for gradients in SQL, and two ML demos #269).Context.prepare(sql)plans a query once, and then runs it many times on new data withprepared(name=dataset). It covers queries that take gradients too, and it reuses their plans between calls (Plan a query once and run it many times: Context.prepare #287, Reuse gradient plans and reduce Dataset registration overhead #301).Context(threads=n)sets how many threads DataFusion uses. It overrides the setting on a copy of the config you pass in, and leaves your config unchanged (Reuse gradient plans and reduce Dataset registration overhead #301).Faster
Changes in behavior
datafusion, move to 55.0.0.Fixes
noleapcalendar keep their calendar date in SQL (fix: keep noleap dates on their calendar date in SQL #309).Docs
🤖 Generated with Claude Code
https://claude.ai/code/session_01NyN7idACeDMh94qrnFMu6h
Summary by CodeRabbit