[dependabot]

Bumps the python group with 53 updates in the / directory:

Package	From	To
pandas	3.0.6	3.1.0
pyarrow	25.0.1	26.0.0
duckdb	1.5.5	2.0.0
geopandas	1.1.4	1.2.0
shapely	2.1.2	2.2.0
city2graph	1.0.0	1.0.1
torch	2.14.0	2.14.1
torch-geometric	2.8.0.post1	2.8.1
marimo	0.24.2	0.25.2
playwright	1.63.0	1.64.1
pillow	12.3.0	13.0.0
ruff	0.16.8	0.17.2
anyio	4.14.2	4.15.1
certifi	2026.7.22	2026.10.23
charset-normalizer	3.4.9	3.5.2
click	8.4.2	8.5.0
contourpy	1.3.3	1.4.0
coverage	7.15.3	7.17.0
fonttools	4.63.0	4.67.0
idna	3.18	3.20
iniconfig	2.3.0	2.3.1
jedi	0.19.2	0.20.1
joblib	1.5.3	1.6.0
kiwisolver	1.5.0	1.5.1
libpysal	4.15.0	4.15.1
loro	1.13.2	1.17.1
markdown	3.10.3	3.11.2
markupsafe	3.0.3	3.0.4
msgspec	0.21.1	0.22.0
narwhals	2.24.0	2.27.1
networkx	3.6.1	3.7
numpy	2.5.1	2.5.4
orjson	3.11.9	3.13.0
overturemaps	1.0.1	1.0.2
packaging	26.2	26.3
platformdirs	4.11.0	4.13.5
pluggy	1.6.0	1.7.0
pycparser	3.0	3.12
pygments	2.20.0	2.21.0
pymdown-extensions	11.0.1	12.3
pyparsing	3.3.2	3.3.3
pyproj	3.7.2	3.8.0
python-dateutil	2.9.0.post0	3.0.0
pyzmq	27.1.0	27.2.0
soupsieve	2.9.1	3.0.4
starlette	1.3.1	1.10.0
threadpoolctl	3.6.0	3.7.0
tomlkit	0.15.1	0.15.2
tqdm	4.70.0	4.70.1
tzdata	2026.3	2026.5
urllib3	2.7.0	2.8.0
uvicorn	0.52.1	0.54.0
websockets	17.0.1	17.3

Updates pandas from 3.0.6 to 3.1.0

Release notes

Sourced from pandas's releases.

pandas 3.1.0

We are pleased to announce the release of pandas 3.1.0. This release includes some new features, bug fixes, and performance improvements. We recommend that all users upgrade to this version.

Highlights include:

• New DataFrame.select() and updated DataFrame.filter() method
• Faster pd.read_csv(), with parallel reads of large files

See the detailed release notes for a list of all the changes.

The release can be installed from PyPI:

python -m pip install --upgrade pandas

Or from conda-forge

conda install -c conda-forge pandas

Please report any issues with the release on the pandas issue tracker.

Thanks to all the contributors who made this release possible.

pandas 3.1.0rc0

We are pleased to announce a release candidate for pandas 3.1.0.

See the whatsnew for a list of all the changes.

The release can be installed from PyPI:

python -m pip install --upgrade --pre pandas==3.1.0rc0

Or from conda-forge

conda install -c conda-forge/label/pandas_rc pandas==3.1.0rc0

Please report any issues with the release candidate on the pandas issue tracker.

Thanks to all the contributors who made this release possible.

Commits

• c6f4676 RLS: 3.1.0
• dd69475 Backport PR #71468 on branch 3.1.x (DOC: 3.1 whatsnew: add some structure to ...
• 4ccc19a Backport PR #71465 on branch 3.1.x (DOC: update announce (contributor listing...
• 24dae63 [backport 3.1.x] DOC: cleanup 3.1.0 whatsnew (#71408) (#71466)
• 8519ca6 [backport 3.1.x] CI: temporary pin dateutil to <3 (#71461) (#71464)
• 2126193 Backport PR #71387 on branch 3.1.x (BUG: fix string regex operations with $ e...
• 4700f5d Backport PR #71442 on branch 3.1.x (BUG: parallel read_csv raising released m...
• 08efb14 Backport PR #71456 on branch 3.1.x (TST: Fix test_reduce_array until reduct...
• ee8dcb7 Backport PR #71429 on branch 3.1.x (DOC: add dedicated 3.1 whatsnew section f...
• d93e32c Backport PR #71383 on branch 3.1.x (BUG: fix string regex operations with Pyt...
• Additional commits viewable in compare view

Updates pyarrow from 25.0.1 to 26.0.0

Release notes

Sourced from pyarrow's releases.

Apache Arrow 26.0.0

Release Notes URL: https://arrow.apache.org/release/26.0.0.html

Apache Arrow 26.0.0 RC0

Release Notes: Release Candidate: 26.0.0 RC0

Commits

• 4461671 MINOR: [Release] Update versions for 26.0.0
• 9d03a26 MINOR: [Release] Update .deb/.rpm changelogs for 26.0.0
• 5ce3c95 MINOR: [Release] Update CHANGELOG.md for 26.0.0
• 28e6c67 GH-51687: [C++] Skipt tests that require threads if ARROW_ENABLE_THREADING=OF...
• 12409f8 GH-51682: [CI][C++] Add liburiparser-dev to system dependency image (#51683)
• 1deceed GH-51675: [Python][Packaging] Apply upstream fix to UriParser from vcpkg and ...
• cca44b5 GH-51638: [CI][C++][Parquet] Update expected JSON output in parquet-reader-te...
• 5183e21 GH-46901: [C++][Compute] Add remainder and modulo kernels (#48914)
• f0d043c MINOR: [Docs][C++] Fix doxygenfile path for arrow/flight/transport.h (#51677)
• 17be9c7 GH-51652: [Python] Add NumPy StringDType to Arrow conversion (#51157)
• Additional commits viewable in compare view

Updates duckdb from 1.5.5 to 2.0.0

Release notes

Sourced from duckdb's releases.

DuckDB v2.0.0

duckdb-python 2.0.0 is the Python package for DuckDB 2.0. It supports Python >= 3.11 and drops Python 3.10. From Python 3.12 on, one wheel per platform serves every Python version. The binding layer was rewritten on nanobind, query results stream instead of being collected up front, the relational API was rebuilt on SQL text, and the way DataFrames, Arrow data and polars frames are read was replaced. Many long-standing bugs are fixed, and moving data in and out of Python is much faster. The engine is DuckDB 2.0.0; see its release notes for what changed in SQL and storage.

Highlights

• Moving data is much faster. Numeric exports to numpy, pandas, Arrow and polars take a quarter to two fifths of the 1.5 time; in-memory Arrow, polars and pandas 3 inputs are scanned on all threads, 8x to 20x faster; Arrow UDFs run about 20x faster; draining an Arrow reader on 16 threads takes an eighth of the time. (#693, #697, #700)
• Results stream. con.execute() starts the query and you read rows as you fetch them; con.complete() and con.abort() finish or cancel an open result, and an open result or Arrow reader owns the connection until it is done.
• The relational API is rebuilt on SQL text. A relation is a lazy query, bound at every step so its schema is known; rel.sql_query() shows exactly what will run. Many new methods and expression helpers (col, lit, when, window functions). One semantic change: a plain string in an expression is a value, not a column name. (#697)
• A new Appender for fast bulk inserts: with con.appender("t") as a: a.append_row(...), under a microsecond per row. (#661)
• One stable ABI wheel per platform for Python 3.12 and later, so a new Python release works on day one. Python 3.15 is supported. (#672, #673, #684)
• Arrow export on con.execute() results changed: fetch_arrow_table, fetch_record_batch and arrow(n) are gone; use con.sql(query).to_arrow_table() or .to_arrow_reader(). (#702)
• The experimental Spark API is brought up to date: it is now compatible with PySpark 4.2.0 and 3.5.9 (#686). We would love your input on whether to continue support or to drop it in future releases.

Upgrading from 1.x: what to check

1. Make sure you run Python 3.11 or later.
2. If you call fetch_arrow_table, fetch_record_batch, arrow(n) or pl() on the result of con.execute(), switch to con.sql(query).to_arrow_table(), to_arrow_reader() or pl(). Libraries that call the removed names on an execute result need an update. (#702)
3. If you call fetchone(), fetchmany(), fetch_df_chunk() or execute() on a relation, use rel.limit(n).fetchall(), rel.to_arrow_reader(), rel.materialize(), or run the query through con.execute(). (#697)
4. If you compare a relation column with a plain string, col("s") == "x" now compares with the text x. Wrap column names in col(). (#697)
5. If you rely on import duckdb having opened the default database, note that it happens on the first duckdb.sql() or duckdb.connect() call instead. (#680)
6. If you pass Decimal values with a positive exponent, such as Decimal("1E+2"), they now arrive as the number they denote, 100, where 1.5 produced 1.00. (#654)
7. If you write files with overwrite=False, an existing file is now refused instead of replaced. (#668)
8. If a test or script depends on str(rel) being free, note that it runs the query each time now.

Breaking changes

Arrow export on DB-API results (#702)

Results of con.execute() are now read row by row and are not kept in memory, so they cannot hand their data to Arrow for free. fetch_arrow_table, fetch_record_batch, arrow(rows_per_batch) and pl() on a DB-API result are removed; they had been deprecated. to_arrow_table() and to_arrow_reader() remain, also as duckdb.to_arrow_table() and duckdb.to_arrow_reader(), but they materialize the rows twice, once as the result and once as Arrow, and emit a UserWarning pointing at the relational path, con.sql(query).to_arrow_table(), which we recommend instead. The reader no longer streams: its first batch arrives after the whole query ran. After a partial fetchone or fetchmany the Arrow call returns the remaining rows instead of raising. With preserve_insertion_order off, the row order can differ between calls. duckdb.arrow(n) with an integer is gone; duckdb.arrow(arrow_object) still builds a relation.

# 1.5
table = con.execute("SELECT * FROM t").fetch_arrow_table()
reader = con.execute("SELECT * FROM t").fetch_record_batch(10_000)
2.0
table = con.sql("SELECT * FROM t").to_arrow_table()
reader = con.sql("SELECT * FROM t").to_arrow_reader(10_000)

Relations are SQL text (#697)

The relational API is a query building API, not an execution API. A relation is a piece of SQL text with a schema bound once when it is built, so every step knows its columns and types. It runs nothing until you ask for data, and it runs again on every read; rel.sql_query() returns exactly the text that runs, as one SELECT per group of steps and a common table expression where the order of clauses breaks. The engine's binder decides every column name and type; nothing is renamed by the client.

What you are most likely to notice:

• A bare string operand of an expression is a value, not a column. For one release a warning is raised when the string matches a column name; the warning goes in 2.1. A dotted string is not split into parts; pass the parts separately.

... (truncated)

Commits

• 83c7199 Build the dictionary test's indices with a 64 bit arange (#705)
• b12fd55 Build the dictionary test's indices with a 64 bit arange
• 6aa2ce3 Build the Linux and macOS wheels with clang and ThinLTO (#703)
• 2f0ac79 Pin submodule to v2.0.0 hash
• f7826ad Fix wrong values and lost parallelism in the pandas, numpy and Arrow input sc...
• 2f2a0dd Bring back to_arrow_table and to_arrow_reader on DB-API results over a re-fee...
• 6086a15 Keep the macOS wheels native, like the engine's release
• 07ee3e2 Build the Linux and macOS wheels with clang and ThinLTO
• a2c151f Register the linked extensions when the module loads, not only on import (#701)
• dd2a3e9 Read numpy masks a word at a time
• Additional commits viewable in compare view

Updates geopandas from 1.1.4 to 1.2.0

Release notes

Sourced from geopandas's releases.

v1.2.0

What's Changed

New features and improvements:

• Static plotting in GeoPandas has been re-implemented. The new tooling shall be mostly backwards compatible while bringing a set of enhancements:

  • Categorical plotting now allows dict-based mapping of categories to colors and other styles.
  • Legends now reflect the geometry type and complete set of styles, not just a color.
  • Plotting of continuous variables with a classification scheme can now produce a colorbar instead of a categorical legend when using colorbar=True within legend_kwds.
  • vmin and vmax keywords are supported when scheme is defined, not ignored.
  • When using categorical colormap with categorical plots, colors are now assigned in order to optimize the contrast of the resulting map.
  • Colorbar size is automatically adjusted to be approximately the size of the plot.
  • Colorbar now indicates extensions if the vmin and/or vmax clip the values.
  • Legends are created automatically by matplotlib. This also means that you can create a single legend from a layered map by calling ax.legend() directly.
  • GeoPandas now labels axes using the information on units from CRS. This can be controlled via the add_labels keyword.
  • Static plots can include a basemap defined by the tiles keyword, exposing contextily.add_basemap from within the plot method. Requires contextily package to be installed.
  • scheme keyword can take a value "greedy" resulting in a greedy (topological) coloring where no two adjacent polygons share a color. Requires mapclassify and libpysal packages.

  The related refactoring of the plotting infrastructure has changed the internal representation of the figure generated by GeoPandas. This can be considered a breaking change if you rely the composition of collections, as that has changed. A number of plotting-related bugs has been fixed along as a side effect. All changes are reflected in an updated user guide. (#3708, #3728, #3748, #3752, #3753, #3763, #3767, #3689).

• GeoParquet improvements:

  • The upcoming GeoParquet 2.0 specification is now supported for both writing and reading (#3632).
  • read_parquet now support direct reading from HTTP/HTTPS protocols (#3699).
  • GeoDataFrame.to_parquet and read_parquet will now write and read attrs respectively (#3597).

• Added geopandas.make_grid and GeoSeries.make_grid to generate a square or hexagonal grids covering the input geometry (#2062).

• Added geopandas.read_file_info to expose information about file layers, including layer metadata (this is an alias for pyogrio.read_info) (#3759).

• Add grid_size parameter to union, difference, symmetric_difference and intersection (#3593).

• Document how to access the feature ID (fid) as the index when reading files

... (truncated)

Changelog

Sourced from geopandas's changelog.

Version 1.2.0 (September 28, 2026)

New features and improvements:

• Static plotting in GeoPandas has been re-implemented. The new tooling shall be mostly backwards compatible while bringing a set of enhancements:

  • Categorical plotting now allows dict-based mapping of categories to colors and other styles.
  • Legends now reflect the geometry type and complete set of styles, not just a color.
  • Plotting of continuous variables with a classification scheme can now produce a colorbar instead of a categorical legend when using colorbar=True within legend_kwds.
  • vmin and vmax keywords are supported when scheme is defined, not ignored.
  • When using categorical colormap with categorical plots, colors are now assigned in order to optimize the contrast of the resulting map.
  • Colorbar size is automatically adjusted to be approximately the size of the plot.
  • Colorbar now indicates extensions if the vmin and/or vmax clip the values.
  • Legends are created automatically by matplotlib. This also means that you can create a single legend from a layered map by calling ax.legend() directly.
  • GeoPandas now labels axes using the information on units from CRS. This can be controlled via the add_labels keyword.
  • Static plots can include a basemap defined by the tiles keyword, exposing contextily.add_basemap from within the plot method. Requires contextily package to be installed.
  • scheme keyword can take a value "greedy" resulting in a greedy (topological) coloring where no two adjacent polygons share a color. Requires mapclassify and libpysal packages.

  The related refactoring of the plotting infrastructure has changed the internal representation of the figure generated by GeoPandas. This can be considered a breaking change if you rely the composition of collections, as that has changed. A number of plotting-related bugs has been fixed along as a side effect. All changes are reflected in an updated user guide. (#3708, #3728, #3748, #3752, #3753, #3763, #3767, #3689).

• GeoParquet improvements:

  • The upcoming GeoParquet 2.0 specification is now supported for both writing and reading (#3632).
  • read_parquet now support direct reading from HTTP/HTTPS protocols (#3699).
  • GeoDataFrame.to_parquet and read_parquet will now write and read attrs respectively (#3597).

• Added geopandas.make_grid and GeoSeries.make_grid to generate a square or hexagonal grids covering the input geometry (#2062).

• Added geopandas.read_file_info to expose information about file layers, including layer metadata (this is an alias for pyogrio.read_info) (#3759).

• Add grid_size parameter to union, difference, symmetric_difference and intersection (#3593).

• Document how to access the feature ID (fid) as the index when reading files with read_file and the pyogrio engine, using fid_as_index=True (#3829).

... (truncated)

Commits

• 199fc53 RLS: v1.2.0
• 0429cd1 DOC/RLS: update changelog for 1.2 release (#3864)
• 3e07c90 BUG: always use int64 for the explode parts index level (#3871)
• 85f3654 TST: use fixed byte-order in to_wkb tests (for big endian machine) (#3874)
• 6d49816 CI: use nightly wheels for shapely in the dev build (#3872)
• 69fb61f ENH: expose read_file_info as toplevel function (#3759)
• cc8e4e8 TST: filterwarning cleanup (#3867)
• 6302753 ENH: support reading GeoParquet 2.0 files with bbox filter (#3847)
• c0e3867 ENH: add new make_grid() function (equivalent of st_make_grid) (#2197)
• 9e58ea1 DOC: consolidate contributing guides (#3859)
• Additional commits viewable in compare view

Updates shapely from 2.1.2 to 2.2.0

Release notes

Sourced from shapely's releases.

2.2.0

Shapely 2.2.0 is a feature release with various new functions, improvements and bug fixes. Highlights include performance improvements for scalar attribute and method calls, and the new minimum_width, coverage_clean, split, transform_coordseq, and get_segments top-level functions.

Shapely 2.2 supports Python >= 3.11, and binary wheels on PyPI include GEOS 3.14 and are now also provided for the Windows ARM64 platform. Wheels are available for Python 3.15.

For a full changelog, see https://shapely.readthedocs.io/en/stable/release/2.x.html#version-2-2-0

2.2.0rc1

Shapely 2.2.0 is a feature release with various new functions, improvements and bug fixes. This is a first release candidate.

For a full changelog, see https://shapely.readthedocs.io/en/latest/release/2.x.html#version-2-2-0

Changelog

Sourced from shapely's changelog.

2.2.0 (2026-10-07)

Shapely 2.2.0 is a feature release with various new functions, improvements and bug fixes. Highlights include performance improvements for scalar attribute and method calls, and the new minimum_width, coverage_clean, split, transform_coordseq, and get_segments top-level functions.

Shapely 2.2 supports Python >= 3.11, and binary wheels on PyPI include GEOS 3.14 and are now also provided for the Windows ARM64 platform. Wheels are available for Python 3.15.

For a full changelog, see https://shapely.readthedocs.io/en/latest/release/2.x.html#version-2-2-0

Commits

• 229c71f RLS: 2.2.0 (#2526)
• 702d8fd CLN: avoid warnings about discarding const qualifier (#2525)
• 0e92040 DOC: update tranform section in the manual (#2518)
• 677f305 MAINT: avoid meson-python 0.22 + DOC: recommend install c-compiler in conda d...
• 68ff25b Use DeprecationWarning instead of FutureWarning for shapely.ops.transform() d...
• 2590a0c BUG: fix empty / non-array cases for transform (#2523)
• 95008b2 ENH: support array input for transform_coordseq (consistent with transform) (...
• e2fe07b docs: Move changelog entry to docs
• c847bb7 fix: Forward **kwargs in line_interpolate_point and line_locate_point (#2480)
• 30c0a51 MAINT: move versioneer config to pyproject.toml; remove setup.cfg (#2511)
• Additional commits viewable in compare view

Updates city2graph from 1.0.0 to 1.0.1

Release notes

Sourced from city2graph's releases.

v1.0.1

What's Changed

Bug fixes

• Fixed knn_graph() self-loops when points share coordinates or snap to the same network node. Thanks to @​pablo-legerensomolinos. (#232)

Documentation

• Published the City2Graph paper and updated citation information and the sitemap. (#226, #227, #228)
• Fixed documentation rendering and reduced notebook page sizes. (#236)
• Added Bing site ownership verification. (#238)

Maintenance

• Updated dependencies, including Tornado from 6.5.7 to 6.5.8, through automated updates by @​dependabot[bot]. (#230, #233, #234, #235, #237)
• Prepared the v1.0.1 release. (#239)

Documentation and release updates were contributed by @​yu-ta-sato.

New Contributors

• @​pablo-legerensomolinos made their first contribution in #232. Thank you!

Full changelog: v1.0.0 → v1.0.1

Changelog

Sourced from city2graph's changelog.

1.0.1 (2026-10-22)

Added

• Added the cu132 extra for CUDA 13.2 installations, with matching official PyTorch and torchvision wheel indexes.

Changed

• Refreshed runtime, development, and documentation dependencies, including DuckDB 2.0, pandas 3.1, GeoPandas 1.2, PyTorch 2.14.1, torchvision 0.29.1, and PyG 2.8.1. CUDA 12.8 remains on PyTorch 2.11.0 and torchvision 0.26.0 because newer CUDA 12.8 wheels are not published.
• Updated the Ruff, mypy, and numpydoc pre-commit hooks, and explicitly declared the documentation's autorefs and PyMdown Extensions dependencies.
• Retained upstream-constrained transitive dependencies: GitPython 3.1.62 and tzdata 2025.3 (mkdocs-rss-plugin), mpmath 1.3.0 (SymPy), and pyee 14.0.0 (Playwright).
• Updated package and citation metadata for v1.0.1.

Fixed

• Fixed knn_graph() creating self-loops and omitting real neighbours when points share coordinates or snap to the same network node. Both coordinate-based and network-based searches now exclude each point by its index.
• Fixed documentation maths rendering and code-copy controls after instant navigation, stabilised the version selector, and avoided duplicate analytics page views.

Documentation

• Updated MathJax to 4.1.3, documented explicit CPU and CUDA wheel selection for pip, and aligned Dependabot's documentation group with the MkDocs toolchain.
• Reduced generated notebook page sizes by removing unused widget state and externalising embedded assets, and added browser smoke tests for documentation interactions.
• Added the published article page and discovery metadata, refreshed citation guidance, and added Bing site ownership verification.

Commits

• 8cffe2d Prepare v1.0.1 release (#239)
• d387266 Bump the uv group across 1 directory with 5 updates (#237)
• c935123 Add Bing site ownership verification (#238)
• 26a5e08 Fix documentation rendering and reduce notebook page payloads (#236)
• 42430fe Bump the uv group across 1 directory with 2 updates (#235)
• e47eac5 Bump the uv group across 1 directory with 3 updates
• 19e01ff Bump the uv group across 1 directory with 2 updates (#233)
• 8c34b04 Fix knn_graph self-loops with coincident points and shared network nodes
• a6b1c7c Bump tornado from 6.5.7 to 6.5.8 in the uv group across 1 directory
• 1fc5321 Update citation info
• Additional commits viewable in compare view

Updates torch from 2.14.0 to 2.14.1

Release notes

Sourced from torch's releases.

PyTorch 2.14.1 Release

This release is meant to fix the following regressions and silent correctness issues:

Silent correctness fixes

• Fix incorrect torch.linalg.lstsq solutions on MPS for complex batched underdetermined systems (#196113), fixed by #196128
• Fix non-orthogonal U and inaccurate small singular values from torch.linalg.svd on MPS for rank-deficient and ill-conditioned inputs (#196112), fixed by #196139 and #199063
• Update the CUDA 13.2 Linux binaries to CUDA 13.2.2 (#196351). This NVIDIA update resolves two critical issues that could produce incorrect results (CUDA 13.2.2 release notes):
  • cuBLAS: cublasLtMatmul() could ignore tensor-wide scaling for NVFP4 matrix multiplications (introduced in CUDA 13.2 Update 1)
  • Compiler: failed thread reconvergence could leave stale or corrupted register values in kernels with nested thread divergence (present since CUDA 12.8)

Regression fixes

• Fix torch.linalg.svd, torch.linalg.svdvals and torch.linalg.lstsq failing on MPS with a Metal pipeline-state error for inputs above 8192 elements (#195937), fixed by #195949 and #195950
• Fix internal assert in torch.svd(out=) on MPS for complex inputs (#195822), fixed by #195872

Commits

• 5c48869 [MPS] Fix Jacobi SVD convergence for small columns (#199086)
• 35223a2 [release/2.14] Run the release runner-group reconcile in pytorch/pytorch (#19...
• 41ffbc4 [release-only] Update version to 2.14.1 (#198661)
• e8f0c01 [release/2.14] [CD] Update to CUDA 13.2.2 for Linux binaries (#198644)
• 911665d [MPS] Cherry-pick native SVD fixes into release/2.14 (#198394)
• See full diff in compare view

Updates torch-geometric from 2.8.0.post1 to 2.8.1

Release notes

Sourced from torch-geometric's releases.

PyG 2.8.1: Bug Fixes

What's Changed

• Prepare for 2.8.0.post1 by @​akihironitta in pyg-team/pytorch_geometric#10749
• Cherry-pick: Fix config_store map_annotation on Python 3.14 (read-only args) (#10726) by @​haoweiyu-kumo in pyg-team/pytorch_geometric#10823

Full Changelog: pyg-team/pytorch_geometric@2.8.0...2.8.1

Commits

• 6bc9810 Cherry-pick: Fix config_store map_annotation on Python 3.14 (read-only __args...
• See full diff in compare view

Updates marimo from 0.24.2 to 0.25.2

Release notes

Sourced from marimo's releases.

0.25.2

This release focuses on bug fixes and adds support for Python 3.15 as well as GeoJSON and GeoParquet downloads for geographic data.

⭐️ Highlights

Python 3.15 support

marimo now supports Python 3.15! Create and run reactive notebooks with the latest Python release. #11128

Export geographic data

Geospatial dataframes (GeoPandas, Arrow, and DuckDB geometry) can now be downloaded as GeoJSON or GeoParquet, alongside the existing CSV, JSON, and Parquet options. Also works with mo.ui.table and mo.ui.dataframe.

🚨 Breaking changes

marimo previously attempted to switch matplotlib, Plotly, Bokeh, and HoloViews to a dark theme when the notebook was dark, but the results were often unexpected, especially in saved figures. Plots now use the plotting library’s own settings. If you want dark plots, please set the theme explicitly in your notebook. (#11147)

✨ Enhancements

• Add support for Python 3.15. #11128
• Add Expand all outputs and Clamp all outputs actions. (#11045)
• Navigate directories with breadcrumbs in mo.ui.file_browser. (#11044)
• Show consistent startup progress for regular notebooks and sandboxes. (#11085)
• Build offline WebAssembly exports using Pyodide and PyPI package mirrors. (#11017)
• Add GeoJSON and GeoParquet exports for geographic tables and dataframes. (#11063, #11116, #11129, #11169)
• Improve Python highlighting inside Markdown LaTeX. (#11281)
• Update the LLM model catalog. (#11100, #11183)

🐛 Bug fixes

• Keep exact mo.ui.number values instead of rounding them to the step size. (#11287)
• Accept datetime inputs in date widgets. (#11289)
• Select the first dropdown option when allow_select_none=False and no value is supplied. (#11238)
• Preserve all rows when sorting tables with unequal-length columns. (#11197)
• Preserve dates when grouping table charts by hour. (#11078)
• Export complete geometry text. (#11012)
• Restore Polars text-to-date/time conversion and preserve empty rows when exploding columns. (#11135, #11158)

... (truncated)

Commits

• b0e4023 release: 0.25.2 (#11327)
• 86ef639 test: handle Windows executable casing in ty LSP test (#11329)
• dd19ff0 fix: apply column format to datetime cells in tables (#11326)
• 169d793 fix(table): pad unequal columns for row selection and search (#11325)
• e8436c8 fix: report SQL errors during automatic dataframe conversion (#11247)
• 4123493 fix: detect ty language server on PATH (#11321)
• b284933 fix(deps): require SQLGlot with matching compiled accelerator (#11322)
• bd3be52 feat: support python 3.15 (#11128)
• 5980629 fix(schema): cover generation and preserve brands on Python 3.10 (#11320)
• 89f57bc Generate named frontend type exports (#11303)
• Additional commits viewable in compare view

Updates playwright from 1.63.0 to 1.64.1

Release notes

Sourced from playwright's releases.

v1.64.1

Highlights

microsoft/playwright-python#3283 - chore: roll Node.js to 24.22.1

The bundled Node.js 24.22.1 is a security release that distrusts the compromised eMudhra intermediate CAs. The driver uses Node's CA store for its own TLS connections, for example connect_over_cdp to a wss:// endpoint.

Browser Versions

• Chromium 156.0.8078.4
• Mozilla Firefox 157.0
• WebKit 27.2

This version was also tested against the following stable channels:

• Google Chrome 155
• Microsoft Edge 155

v1.64.0

🧰 WebMCP

New page.webmcp and frame.webmcp give access to the tools that a page registers through the experimental WebMCP browser API, so you can test them like any other part of your app:

browser = chromium.launch(args=["--enable-features=WebMCP"])
page = browser.new_page()
page.goto("https://example.com")
for tool in page.webmcp.tools():
print(tool["name"], tool["description"])
result = page.webmcp.call_tool("add", {"a": 2, "b": 40})
print(result["content"][0]["text"])  # "42"

Playwright MCP also supports WebMCP by default, with page-defined tools offered to the agent as webmcp_<tool>. Pass --no-webmcp to opt out.

playwright-cli exposes them as well:

playwright-cli webmcp-list
playwright-cli webmcp-call search_catalog --params '{"query":"cats"}'

🎬 Better videos

Videos can now be recorded at a custom frame rate, and the decorations for actions are styled with plain CSS:

... (truncated)

Commits

• bcb9deb cherry-pick(#3283): chore: roll Node.js to 24.22.1
• 387a8a9 cherry-pick(#3252): devops: publish to PyPI from GitHub Actions with trusted ...
• 71d2a17 chore: roll Playwright to v1.64.0 (#3251)
• d566fc5 fix(fetch): JSON-stringify NaN and Infinity string data (#3249)
• 2959d19 docs(skills): add a skill to prepare a release (#3250)
• 5313ac2 fix(network): keep blank values in Request.post_data_json for form bodies (#3...
• 9d8c50f chore(codegen): use alphabetical order in the generated API (#3242)
• 478282f chore(codegen): generate the public TypedDicts from the docs (#3247)
• b8b7155 fix: raise Error instead of assert for user input validation (#3245)
• 1de5d37 fix(network): do not add missing headers when looking them up (#3246)
• Additional commits viewable in compare view

Updates pillow from 12.3....

Description has been truncated