feat: add support for Python 3.14 by gforsyth · Pull Request #21540 · rapidsai/cudf

gforsyth · 2026-02-24T20:00:04Z

Description

Contributes to rapidsai/build-planning#205

This PR adds support for Python 3.14.

Notes for Reviewers

This is part of ongoing work to add Python 3.14 support across RAPIDS.
It temporarily introduces a build/test matrix including Python 3.14, from rapidsai/shared-workflows#508.

A follow-up PR will revert back to pointing at the main branch of shared-workflows once all
RAPIDS repos have added Python 3.14 support.

This will fail until all dependencies have been updated to Python 3.14

CI here is expected to fail until all of this project's upstream dependencies support Python 3.14.

This can be merged whenever all CI jobs are passing.

copy-pr-bot · 2026-02-24T20:00:08Z

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

gforsyth · 2026-02-24T20:10:24Z

/ok to test

xref #21540 xref rapidsai/build-planning#205 xref rapidsai/dask-cuda#1629 We _should_ be able to merge in Python 3.14 support PRs without admin-merges because of limited APIs support, but the `numba` dependency here needs to be bumped for Python 3.14 testing in `dask-cuda` and others.

gforsyth · 2026-02-26T21:53:57Z

/ok to test

gforsyth · 2026-02-26T23:32:47Z

/ok to test

gforsyth · 2026-02-27T16:27:28Z

/ok to test

gforsyth · 2026-02-27T18:42:32Z

/ok to test

gforsyth · 2026-02-27T19:47:11Z

/ok to test

…hey shouldn't

gforsyth · 2026-02-27T22:45:30Z

/ok to test

gforsyth · 2026-02-27T22:46:47Z

ci/cudf_pandas_scripts/run_tests.sh

+version_lte() {
+  [ "$1" = "$(echo -e "$1\n$2" | sort -V | head -n1)" ]
+}
+
+if version_lte "${RAPIDS_PY_VERSION}" "3.13"; then
+    for version in "${versions[@]}"; do
+        echo "Installing pandas version: ${version}"
+        # This loop tests cudf.pandas compatibility with older pandas-numpy versions,
+        # requiring numpy<2. cupy>=14 dropped support for numpy<2, so we explicitly
+        # downgrade cupy here to avoid an import failure when cupy tries
+        # to load against the older numpy.


Python 3.14 requires numpy>=2 so skipping this check for Python >3.13

In the "else" block please print something like "Python ${RAPIDS_PY_VERSION} detected, which is newer than 3.13. Skipping cudf.pandas compatibility tests with numpy < 2."

Agree that a print would be helpful, the change is fine though. If this winds up being the only change request feel free to add the print in a follow-up PR fixing the pr.yaml changed files.

gforsyth · 2026-02-27T22:47:29Z

ci/test_cudf_polars_polars_tests.sh

 # in versions < 2.12, and trigger a pyparsing DeprecationWarning during collection.
 # See https://github.com/pola-rs/polars/pull/25854
-sed -i 's/^pydantic>=2.0.0.*/pydantic>=2.0.0,<2.12.0/' polars/py-polars/requirements-dev.txt
+sed -i 's/^pydantic>=2.0.0.*/pydantic>=2.0.0,<2.13.0/' polars/py-polars/requirements-dev.txt


I think it was pydantic that caused failures from pyo3 when compiling with Python 3.14, so bumping that bound as 2.12.0 apparently has 3.14 support

I'm a bit confused here. The comment above specifically calls out versions < 2.12 so I am hesitant to change this unless we have established this is safe to change. Probably requires tracing through the linked PR and figure out what versions of what packages are affected.

Ok, this is a hard blocker for Python 3.14 support

I'm going to defer to @Matt711 for this question since he would have added this exact filter.

Okay a couple of things for more context:

There was a regression in pydantic 2.12.0 (Accelerate stable-segmented-sort with CUB segmented sort #12347) that broke pyiceberg tests in polars. This was

Also there's a deprecation warning: PydanticDeprecatedSince212 (doesn't exist in < 2.12) that the pyiceberg tests we're using to filter with.

For 1., this was fixed in pydantic 2.12.1 which we would pick up in this new pinning (so that's 👍🏾 )
For 2., we don't need filter the warning out in https://github.com/rapidsai/cudf/pull/21540/changes#diff-7132576ea4ccfd76284de2d8aced9ed5a3e4f2d997c27b7b5ecd2447702361d8R45-R46 anymore now that we'd pick up 2.12.1.

This shouldn't block this PR, but ideally the pydantic pinning would be >=2.12.1,<2.13.0 and we can remove the code that filters the PydanticDeprecatedSince212 in this script.

Or actually the pinning would be >=2.12.1 and with the upper bound removed entirely

Ok -- I'll merge this in so we get the test filters fixed in CI, then I can put up a follow-up to handle the warning removal, bump the pydantic pinning, and fix up the changed-files filters

gforsyth · 2026-02-27T22:47:53Z

ci/test_python_cudf.sh

+version_gte() {
+  [ "$2" = "$(echo -e "$2\n$1" | sort -V | head -n1)" ]
+}
+
 # If the RAPIDS_PY_VERSION is 3.13, set CUDF_TEST_COPY_ON_WRITE to '1' to enable copy-on-write tests.
-if [[ "${RAPIDS_PY_VERSION}" == "3.13" ]]; then
+if version_gte "${RAPIDS_PY_VERSION}" "3.13"; then


Assuming we want to run CoW tests for >=Python 3.13, not just 3.13

I am guessing this was just meant to be "latest" for partial coverage. @vyasr?

It was just meant for partial coverage, yes. There's no harm in making this a gte, though. It's fine to change what percentage of the matrix runs with CoW on or off as long as we have some coverage in both cases. This branching will be removed before we start to support the next Python version (3.15) since by then cudf will require pandas 3.0 where copy-on-write is the only supported behavior.

bdice · 2026-03-02T15:25:18Z

.github/workflows/pr.yaml

-          - 'python/cudf_polars/**'
-        dask_cudf_only:
-          - 'python/dask_cudf/**'
+        # cpp_only:


Reminder to revert this and any other temporary changes.

I don't think these should be reverted. The filters are wrong and are skipping portions of the test suite when they shouldn't.

Yeah we discussed this on Slack, we're skipping tests that we shouldn't. We should be either removing these filters or correcting them.

This is frustrating. Can I at least add the cpp_only one back?

It was added back in a working form in #21622

Thanks Gil.

bdice · 2026-03-02T15:26:21Z

ci/cudf_pandas_scripts/run_tests.sh

+version_lte() {
+  [ "$1" = "$(echo -e "$1\n$2" | sort -V | head -n1)" ]
+}
+
+if version_lte "${RAPIDS_PY_VERSION}" "3.13"; then
+    for version in "${versions[@]}"; do
+        echo "Installing pandas version: ${version}"
+        # This loop tests cudf.pandas compatibility with older pandas-numpy versions,
+        # requiring numpy<2. cupy>=14 dropped support for numpy<2, so we explicitly
+        # downgrade cupy here to avoid an import failure when cupy tries
+        # to load against the older numpy.


In the "else" block please print something like "Python ${RAPIDS_PY_VERSION} detected, which is newer than 3.13. Skipping cudf.pandas compatibility tests with numpy < 2."

bdice · 2026-03-02T15:28:15Z

ci/test_cudf_polars_polars_tests.sh

 # in versions < 2.12, and trigger a pyparsing DeprecationWarning during collection.
 # See https://github.com/pola-rs/polars/pull/25854
-sed -i 's/^pydantic>=2.0.0.*/pydantic>=2.0.0,<2.12.0/' polars/py-polars/requirements-dev.txt
+sed -i 's/^pydantic>=2.0.0.*/pydantic>=2.0.0,<2.13.0/' polars/py-polars/requirements-dev.txt


I'm a bit confused here. The comment above specifically calls out versions < 2.12 so I am hesitant to change this unless we have established this is safe to change. Probably requires tracing through the linked PR and figure out what versions of what packages are affected.

bdice · 2026-03-02T15:28:46Z

ci/test_python_cudf.sh

+version_gte() {
+  [ "$2" = "$(echo -e "$2\n$1" | sort -V | head -n1)" ]
+}
+
 # If the RAPIDS_PY_VERSION is 3.13, set CUDF_TEST_COPY_ON_WRITE to '1' to enable copy-on-write tests.
-if [[ "${RAPIDS_PY_VERSION}" == "3.13" ]]; then
+if version_gte "${RAPIDS_PY_VERSION}" "3.13"; then


I am guessing this was just meant to be "latest" for partial coverage. @vyasr?

vyasr

We should wait for @Matt711 to confirm if the pydantic change is OK. The rest looks fine so I'm approving from my end.

vyasr · 2026-03-02T17:30:53Z

.github/workflows/pr.yaml

-          - 'python/cudf_polars/**'
-        dask_cudf_only:
-          - 'python/dask_cudf/**'
+        # cpp_only:


Yeah we discussed this on Slack, we're skipping tests that we shouldn't. We should be either removing these filters or correcting them.

vyasr · 2026-03-02T17:34:34Z

ci/cudf_pandas_scripts/run_tests.sh

+version_lte() {
+  [ "$1" = "$(echo -e "$1\n$2" | sort -V | head -n1)" ]
+}
+
+if version_lte "${RAPIDS_PY_VERSION}" "3.13"; then
+    for version in "${versions[@]}"; do
+        echo "Installing pandas version: ${version}"
+        # This loop tests cudf.pandas compatibility with older pandas-numpy versions,
+        # requiring numpy<2. cupy>=14 dropped support for numpy<2, so we explicitly
+        # downgrade cupy here to avoid an import failure when cupy tries
+        # to load against the older numpy.


Agree that a print would be helpful, the change is fine though. If this winds up being the only change request feel free to add the print in a follow-up PR fixing the pr.yaml changed files.

vyasr · 2026-03-02T17:38:07Z

ci/test_python_cudf.sh

+version_gte() {
+  [ "$2" = "$(echo -e "$2\n$1" | sort -V | head -n1)" ]
+}
+
 # If the RAPIDS_PY_VERSION is 3.13, set CUDF_TEST_COPY_ON_WRITE to '1' to enable copy-on-write tests.
-if [[ "${RAPIDS_PY_VERSION}" == "3.13" ]]; then
+if version_gte "${RAPIDS_PY_VERSION}" "3.13"; then


It was just meant for partial coverage, yes. There's no harm in making this a gte, though. It's fine to change what percentage of the matrix runs with CoW on or off as long as we have some coverage in both cases. This branching will be removed before we start to support the next Python version (3.15) since by then cudf will require pandas 3.0 where copy-on-write is the only supported behavior.

vyasr · 2026-03-02T17:38:28Z

ci/test_cudf_polars_polars_tests.sh

 # in versions < 2.12, and trigger a pyparsing DeprecationWarning during collection.
 # See https://github.com/pola-rs/polars/pull/25854
-sed -i 's/^pydantic>=2.0.0.*/pydantic>=2.0.0,<2.12.0/' polars/py-polars/requirements-dev.txt
+sed -i 's/^pydantic>=2.0.0.*/pydantic>=2.0.0,<2.13.0/' polars/py-polars/requirements-dev.txt


I'm going to defer to @Matt711 for this question since he would have added this exact filter.

gforsyth · 2026-03-02T18:14:00Z

/merge

This reverts commit edf65a0.

- **fix(test_cudf_polars): set new pydantic lower-bound, remove warning filtering** - **chore(cudf-pandas): print message when skipping pandas.cudf compat tests** - **refactor(changed-files): create negation groups for job skipping** Cleanup and followup of some changes made in #21540 Authors: - Gil Forsyth (https://github.com/gforsyth) - Matthew Murray (https://github.com/Matt711) Approvers: - Kyle Edwards (https://github.com/KyleFromNVIDIA) - James Lamb (https://github.com/jameslamb) URL: #21622

## Description Scanning https://anaconda.org/channels/conda-forge/packages/librdkafka, it appears that 2.11.1 is the first version to support Python 3.14. I think this is the cause of the devcontainer conda failures e.g. https://github.com/rapidsai/cudf/actions/runs/23554430068/job/68577154587?pr=21921 (below). So this PR bumps `librdkafka` and follows our similar, narrow pinning so we support Python 3.14 as well. However, the [python-confluent-kafka recipe](https://github.com/conda-forge/python-confluent-kafka-feedstock/blob/main/recipe/meta.yaml#L38) specifies `- librdkafka >={{ version }}` so we're effectively also pinning `python-confluent-kafka` even though it was unpinned on our side in #21540 ``` The following packages are incompatible ├─ librdkafka >=2.8.0,<2.9.0 * is requested and can be installed; ├─ python-confluent-kafka =* * is installable with the potential options │ ├─ python-confluent-kafka 2.1.0 would require │ │ └─ librdkafka >=2.1.0,<2.2.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.1.1 would require │ │ └─ librdkafka >=2.1.1,<2.2.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.10.0 would require │ │ └─ librdkafka >=2.10.0,<2.11.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.10.1 would require │ │ └─ librdkafka >=2.10.1,<2.11.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.11.0 would require │ │ └─ librdkafka >=2.11.0,<2.12.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.11.1 would require │ │ └─ librdkafka >=2.11.1,<2.12.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.12.2 would require │ │ └─ librdkafka >=2.12.1,<2.13.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.13.0 would require │ │ └─ librdkafka >=2.13.0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.13.2 would require │ │ └─ librdkafka >=2.13.2 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.2.0 would require │ │ └─ librdkafka >=2.2.0,<2.3.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.3.0 would require │ │ └─ librdkafka >=2.3.0,<2.4.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.4.0 would require │ │ └─ librdkafka >=2.4.0,<2.5.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.5.0 would require │ │ └─ librdkafka >=2.5.0,<2.6.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.5.3 would require │ │ └─ librdkafka >=2.5.3,<2.6.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.6.0 would require │ │ └─ librdkafka >=2.6.0,<2.7.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.6.1 would require │ │ └─ librdkafka >=2.6.1,<2.7.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 2.8.0 would require │ │ └─ python_abi =3.10 *_cp310, which can be installed; │ ├─ python-confluent-kafka 2.8.0 would require │ │ └─ python_abi =3.11 *_cp311, which can be installed; │ ├─ python-confluent-kafka 2.8.0 would require │ │ └─ python_abi =3.12 *_cp312, which can be installed; │ ├─ python-confluent-kafka 2.8.0 would require │ │ └─ python_abi =3.13 *_cp313, which can be installed; │ ├─ python-confluent-kafka 2.8.0 would require │ │ └─ python_abi =3.9 *_cp39, which can be installed; │ ├─ python-confluent-kafka [0.11.0|0.11.4|0.9.4] would require │ │ └─ librdkafka [==0.9.4 *|>=0.9.4,<0.9.5.0a0 *], which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 0.11.4 would require │ │ └─ librdkafka >=0.11.0,<0.12 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 0.11.4 would require │ │ └─ librdkafka >=0.11.5,<0.11.6.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 1.0.1 would require │ │ └─ librdkafka >=1.0.1,<1.0.2.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka [1.1.0|1.3.0] would require │ │ └─ python_abi =* *_cp27mu, which can be installed; │ ├─ python-confluent-kafka [1.1.0|1.3.0] would require │ │ └─ python_abi =* *_cp36m, which can be installed; │ ├─ python-confluent-kafka [1.1.0|1.3.0] would require │ │ └─ python_abi =* *_cp37m, which can be installed; │ ├─ python-confluent-kafka [1.1.0|1.3.0] would require │ │ └─ python_abi =* *_cp38, which can be installed; │ ├─ python-confluent-kafka 1.3.0 would require │ │ └─ librdkafka >=1.4.0,<1.5.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka [1.3.0|1.5.0] would require │ │ └─ librdkafka >=1.5.0,<1.6.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka [1.5.0|1.6.0] would require │ │ └─ librdkafka >=1.6.0,<1.7.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 1.7.0 would require │ │ └─ librdkafka >=1.7.0,<1.8.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 1.8.2 would require │ │ └─ librdkafka >=1.8.2.0inf.2,<1.9.0a0 *, which conflicts with any installable versions previously reported; │ ├─ python-confluent-kafka 1.9.0 would require │ │ └─ librdkafka >=1.9.1,<1.10.0a0 *, which conflicts with any installable versions previously reported; │ └─ python-confluent-kafka [1.9.0|1.9.2] would require │ └─ librdkafka >=1.9.2,<1.10.0a0 *, which conflicts with any installable versions previously reported; └─ python =3.14 * is not installable because there are no viable options ├─ python [3.14.0|3.14.1|3.14.2|3.14.3] would require │ └─ python_abi =3.14 *_cp314, which conflicts with any installable versions previously reported; ├─ python [3.14.0|3.14.1|3.14.2|3.14.3] would require │ └─ python_abi =3.14 *_cp314t, which conflicts with any installable versions previously reported; └─ python [3.14.0rc1|3.14.0rc2|3.14.0rc3] would require └─ _python_rc =* *, which does not exist (perhaps a missing channel). ``` ## Checklist - [ ] I am familiar with the [Contributing Guidelines](https://github.com/rapidsai/cudf/blob/HEAD/CONTRIBUTING.md). - [ ] New or existing tests cover these changes. - [ ] The documentation is up to date with these changes. Admin merged per eng request

gforsyth added improvement Improvement / enhancement to an existing function non-breaking Non-breaking change labels Feb 24, 2026

github-actions bot assigned gforsyth Feb 24, 2026

github-actions bot added Python Affects Python cuDF API. cudf.pandas Issues specific to cudf.pandas cudf-polars Issues specific to cudf-polars pylibcudf Issues specific to the pylibcudf package labels Feb 24, 2026

github-project-automation bot added this to cuDF Python Feb 24, 2026

GPUtester moved this to In Progress in cuDF Python Feb 24, 2026

gforsyth mentioned this pull request Feb 24, 2026

Python 3.14 support rapidsai/build-planning#205

Open

38 tasks

gforsyth mentioned this pull request Feb 24, 2026

chore(deps): bump numba ceiling to <0.65.0 #21543

Merged

gforsyth force-pushed the python-3.14 branch 2 times, most recently from d401554 to 9239d76 Compare February 26, 2026 21:53

gforsyth force-pushed the python-3.14 branch from 42b1f7a to 8ffa4c1 Compare February 26, 2026 23:12

gforsyth force-pushed the python-3.14 branch from ad9c93e to 37117a7 Compare February 27, 2026 18:42

gforsyth force-pushed the python-3.14 branch from 7d1898e to d9d0d47 Compare February 27, 2026 19:47

gforsyth added 6 commits February 27, 2026 17:45

feat: add support for Python 3.14

b40ce62

chore(deps): remove python-confluent-kafka pin

4f1ff53

chore: strip out librdkafka version

f8c5183

chore: remove *_only changed-files checks which are skipping jobs t…

5c83281

…hey shouldn't

fix(cudf_pandas_scripts): don't test numpy<2 on Python >=3.14

08a07a4

fix(test): test CoW for Python >=3.13

f35f891

test(deps): bump upper bound of pydantic for Python 3.14 support

b94b323

gforsyth force-pushed the python-3.14 branch from ccb38b7 to b94b323 Compare February 27, 2026 22:45

gforsyth commented Feb 27, 2026

View reviewed changes

gforsyth marked this pull request as ready for review March 2, 2026 14:47

gforsyth requested review from a team as code owners March 2, 2026 14:47

gforsyth requested a review from AyodeAwe March 2, 2026 14:47

bdice reviewed Mar 2, 2026

View reviewed changes

vyasr approved these changes Mar 2, 2026

View reviewed changes

rapids-bot bot merged commit edf65a0 into rapidsai:main Mar 2, 2026
116 checks passed

github-project-automation bot moved this from In Progress to Done in cuDF Python Mar 2, 2026

gforsyth deleted the python-3.14 branch March 2, 2026 18:20

gforsyth mentioned this pull request Mar 2, 2026

chore: followup fixes for Python 3.14 and test filtering #21622

Merged

mroeschke mentioned this pull request Mar 2, 2026

Revert "Bump minimum NumPy dependency to 1.26" #21623

Closed

mroeschke added a commit to mroeschke/cudf that referenced this pull request Mar 3, 2026

Revert "feat: add support for Python 3.14 (rapidsai#21540)"

5a7d2d9

This reverts commit edf65a0.

mroeschke mentioned this pull request Mar 30, 2026

Bump librdkafka to >=2.11.1,<2.12.0 #21951

Merged

3 tasks

Conversation

gforsyth commented Feb 24, 2026

Description

Notes for Reviewers

This will fail until all dependencies have been updated to Python 3.14

Uh oh!

copy-pr-bot bot commented Feb 24, 2026

Uh oh!

gforsyth commented Feb 24, 2026

Uh oh!

gforsyth commented Feb 26, 2026

Uh oh!

gforsyth commented Feb 26, 2026

Uh oh!

gforsyth commented Feb 27, 2026

Uh oh!

gforsyth commented Feb 27, 2026

Uh oh!

gforsyth commented Feb 27, 2026

Uh oh!

gforsyth commented Feb 27, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

vyasr left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

gforsyth commented Mar 2, 2026