Define pandas and polars compatibility testing versions in dependencies.yaml - #22189
Merged
rapids-bot[bot] merged 11 commits intoJun 2, 2026
Merged
Conversation
Matt711
added
improvement
|
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. |
Member
Author
|
/ok to test 81b0f70 |
Matt711
commented
Apr 17, 2026
Comment on lines
+1410
to
+1413
| - matrix: | ||
| polars_compat_version: "1.34" | ||
| packages: | ||
| - polars==1.34.* |
Member
Author
There was a problem hiding this comment.
Why is the polars binary missing for version 1.34?
RAPIDS logger » [04/17/26 13:49:13]
┌───────────────────────────────────────┐
| Running tests for polars 1.34.* |
└───────────────────────────────────────┘
ImportError while loading conftest '/__w/cudf/cudf/python/cudf_polars/tests/conftest.py'.
tests/conftest.py:9: in <module>
import cudf_polars.callback
cudf_polars/__init__.py:14: in <module>
from cudf_polars.callback import execute_with_cudf
cudf_polars/callback.py:20: in <module>
from polars.exceptions import ComputeError, PerformanceWarning
../../venv_polars_1.34/lib/python3.14/site-packages/polars/__init__.py:56: in <module>
from polars import api, exceptions, plugins, selectors
../../venv_polars_1.34/lib/python3.14/site-packages/polars/api.py:6: in <module>
import polars._reexport as pl
../../venv_polars_1.34/lib/python3.14/site-packages/polars/_reexport.py:3: in <module>
from polars.dataframe import DataFrame
../../venv_polars_1.34/lib/python3.14/site-packages/polars/dataframe/__init__.py:1: in <module>
from polars.dataframe.frame import DataFrame
../../venv_polars_1.34/lib/python3.14/site-packages/polars/dataframe/frame.py:32: in <module>
from polars import functions as F
../../venv_polars_1.34/lib/python3.14/site-packages/polars/functions/__init__.py:32: in <module>
from polars.functions.lazy import (
../../venv_polars_1.34/lib/python3.14/site-packages/polars/functions/lazy.py:8: in <module>
import polars.selectors as cs
../../venv_polars_1.34/lib/python3.14/site-packages/polars/selectors.py:33: in <module>
from polars.expr import Expr
../../venv_polars_1.34/lib/python3.14/site-packages/polars/expr/__init__.py:1: in <module>
from polars.expr.expr import Expr
../../venv_polars_1.34/lib/python3.14/site-packages/polars/expr/expr.py:66: in <module>
from polars.meta import thread_pool_size
../../venv_polars_1.34/lib/python3.14/site-packages/polars/meta/__init__.py:3: in <module>
from polars.meta.build import build_info
../../venv_polars_1.34/lib/python3.14/site-packages/polars/meta/build.py:5: in <module>
from polars._utils.polars_version import get_polars_version
../../venv_polars_1.34/lib/python3.14/site-packages/polars/_utils/polars_version.py:9: in <module>
warnings.warn("Polars binary is missing!", stacklevel=2)
E UserWarning: Polars binary is missing!
RAPIDS logger » [04/17/26 13:49:13]
┌──────────────────────────────────────┐
| Tests failed for polars 1.34.* |
└──────────────────────────────────────┘
Member
Author
|
/ok to test ffd9949 |
Matt711
force-pushed
the
imp/ci/pandas-polars-compat
branch
from
April 17, 2026 14:43
ffd9949 to
8443e4b
Compare
Member
Author
|
/ok to test fb4de7a |
3 tasks
Contributor
|
I merged this into #22265 and will continue there. |
3 tasks
rapids-bot Bot
pushed a commit
that referenced
this pull request
Apr 23, 2026
This is a simpler fix for the problem #22189 is aiming to solve that should get us unblocked in the near term. Authors: - Vyas Ramasubramani (https://github.com/vyasr) Approvers: - Bradley Dice (https://github.com/bdice) URL: #22273
shrshi
pushed a commit
to shrshi/cudf
that referenced
this pull request
May 12, 2026
This is a simpler fix for the problem NVIDIA#22189 is aiming to solve that should get us unblocked in the near term. Authors: - Vyas Ramasubramani (https://github.com/vyasr) Approvers: - Bradley Dice (https://github.com/bdice) URL: NVIDIA#22273
Contributor
|
/ok to test f40253e |
Contributor
|
/ok to test 5f7106e |
Contributor
|
/ok to test dc2d077 |
mroeschke
marked this pull request as ready for review
May 28, 2026 23:01
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: 📒 Files selected for processing (2)
🚧 Files skipped from review as they are similar to previous changes (2)
📝 WalkthroughSummary by CodeRabbit
WalkthroughAdds dependency matrix entries for pandas and polars, a CLI to extract matrix values from dependencies.yaml, and refactors CI scripts to run per-compatibility-version tests in isolated virtual environments using generated per-version requirement files. ChangesCompatibility version test infrastructure
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Suggested labels
Suggested reviewers
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 3
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
ci/test_wheel_cudf_polars.sh (1)
36-83:⚠️ Potential issue | 🟠 Major | ⚡ Quick winAdd a GPU availability preflight check before the test loop.
The script runs GPU-dependent tests without an explicit GPU check. Add a fail-fast probe so CI errors are immediate and actionable.
As per coding guidelines, "Verify GPU availability checks before tests".Suggested patch
PASSED=() FAILED=() +if ! nvidia-smi -L >/dev/null 2>&1; then + rapids-logger "GPU not detected; cannot run cudf_polars compatibility tests" + exit 1 +fi + for version in "${VERSIONS[@]}"; do🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@ci/test_wheel_cudf_polars.sh` around lines 36 - 83, Add a GPU availability preflight immediately before the for version in "${VERSIONS[@]}" loop: run a probe (e.g., call nvidia-smi or a small Python/cuda import) and if it fails log an error with rapids-logger and exit non‑zero so CI fails fast; ensure the check references the same environment used by tests (GPU presence for RAPIDS/CUDA) and guard the loop by returning non-zero when no GPUs are detected.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@ci/cudf_pandas_scripts/run_tests.sh`:
- Around line 90-125: The loop that sets up env_name, activates the venv and
installs deps assumes a GPU is present but never verifies it; add a fail-fast
GPU preflight check (e.g., run `nvidia-smi -L` and exit non-zero with a clear
error message) inside the loop after sourcing "${env_name}/bin/activate" and
before the `python -m pytest -p cudf.pandas` invocation so tests abort early on
CI when no GPU is available.
In `@ci/test_wheel_cudf_polars.sh`:
- Around line 48-69: The venv creation/activation and dependency installation
steps (env_name, python -m venv --clear, source "${env_name}/bin/activate", and
rapids-pip-retry install) are currently unguarded so failures are ignored;
modify the script to immediately stop and report an error if any of these steps
fail—either enable errexit (set -e) around the venv/install block or explicitly
check the exit status after python -m venv, after source
"${env_name}/bin/activate", and after rapids-pip-retry install and call a
failure handler (logging via rapids-logger and exit nonzero) so tests never run
in a broken environment; apply the same guarding pattern to the later venv block
referenced for lines 91-103.
In `@ci/utils/get_matrix_values.py`:
- Around line 24-30: The code currently treats the fetched `values` as iterable,
which will iterate characters if a scalar string is returned; update the block
that reads `values = matrix.get(matrix_var)` to validate the type before
iterating: check that `values` is a sequence (e.g., list/tuple or a
collections.abc.Sequence) and explicitly reject strings/bytes, and if it is not
a proper sequence raise a RuntimeError (include `matrix_var` and `file_key` in
the message) instead of silently iterating; then safely return `[str(v) for v in
values]`.
---
Outside diff comments:
In `@ci/test_wheel_cudf_polars.sh`:
- Around line 36-83: Add a GPU availability preflight immediately before the for
version in "${VERSIONS[@]}" loop: run a probe (e.g., call nvidia-smi or a small
Python/cuda import) and if it fails log an error with rapids-logger and exit
non‑zero so CI fails fast; ensure the check references the same environment used
by tests (GPU presence for RAPIDS/CUDA) and guard the loop by returning non-zero
when no GPUs are detected.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: f0b4f98b-c72e-4ebd-ba3f-42c22262e657
📒 Files selected for processing (4)
ci/cudf_pandas_scripts/run_tests.shci/test_wheel_cudf_polars.shci/utils/get_matrix_values.pydependencies.yaml
Comment on lines
90
to
125
| if version_lte "${RAPIDS_PY_VERSION}" "3.13"; then | ||
| for version in "${versions[@]}"; do | ||
| echo "Installing pandas version: ${version}" | ||
| # This loop tests cudf.pandas compatibility with older pandas-numpy versions, | ||
| # requiring numpy<2. cupy>=14 dropped support for numpy<2, so we explicitly | ||
| # downgrade cupy here to avoid an import failure when cupy tries | ||
| # to load against the older numpy. | ||
| set +e | ||
| # We're iterating over a range of versions that might not all have | ||
| # supported wheels, so our best bet for now is to disallow source | ||
| # distributions. | ||
| rapids-pip-retry install "numpy>=1.26,<2.0a0" "pandas==${version}" "cupy-cuda${RAPIDS_CUDA_VERSION%%.*}x<14" --only-binary=:all: | ||
| INSTALL_SUCCESS=$? | ||
| set -e | ||
| if [[ ${INSTALL_SUCCESS} -ne 0 ]]; then | ||
| echo "Failed to install pandas ${version} with numpy<2 for Python ${RAPIDS_PY_VERSION}. Skipping tests for this version." | ||
| continue | ||
| fi | ||
| rapids-logger "Testing cudf.pandas compatibility with pandas ${version}.*" | ||
|
|
||
| # Generate requirements for this pandas compat version. | ||
| # Each entry pins numpy<2 + the specific pandas minor line + the CUDA-appropriate cupy<14. | ||
| # cupy>=14 dropped support for numpy<2 (see https://github.com/cupy/cupy/issues/9709). | ||
| pandas_requirements_txt="pandas-compat-${version}-requirements.txt" | ||
| rapids-dependency-file-generator \ | ||
| --config dependencies.yaml \ | ||
| --file-key test_cudf_pandas_compat \ | ||
| --output requirements \ | ||
| --matrix "cuda=${RAPIDS_CUDA_VERSION};pandas_compat_version=${version}" \ | ||
| > "${pandas_requirements_txt}" | ||
|
|
||
| env_name="venv_pandas_${version}" | ||
| python -m venv --clear "${env_name}" | ||
| # shellcheck disable=SC1090 | ||
| source "${env_name}/bin/activate" | ||
|
|
||
| # notes: | ||
| # | ||
| # * echo to expand wildcard before adding `[test,cudf-pandas-tests]` requires for pip | ||
| # * need to provide --constraint="${PIP_CONSTRAINT}" because that environment variable is | ||
| # ignored if any other --constraint are passed via the CLI | ||
| # | ||
| rapids-pip-retry install \ | ||
| -v \ | ||
| --constraint ./constraints.txt \ | ||
| --constraint "${PIP_CONSTRAINT}" \ | ||
| "$(echo "${CUDF_WHEELHOUSE}"/cudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)[test,cudf-pandas-tests]" \ | ||
| "$(echo "${LIBCUDF_WHEELHOUSE}"/libcudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)" \ | ||
| "$(echo "${PYLIBCUDF_WHEELHOUSE}"/pylibcudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)" \ | ||
| -r "${pandas_requirements_txt}" | ||
|
|
||
| python -m pytest -p cudf.pandas \ |
There was a problem hiding this comment.
Add an explicit GPU preflight check before running compat tests.
This loop assumes CUDA/GPU availability but never verifies it before invoking pytest. Add a fail-fast guard (e.g., nvidia-smi -L) to produce a clear CI error.
Suggested patch
if version_lte "${RAPIDS_PY_VERSION}" "3.13"; then
+ if ! nvidia-smi -L >/dev/null 2>&1; then
+ rapids-logger "GPU not detected; cannot run cudf.pandas compatibility tests"
+ exit 1
+ fi
for version in "${versions[@]}"; do📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| if version_lte "${RAPIDS_PY_VERSION}" "3.13"; then | |
| for version in "${versions[@]}"; do | |
| echo "Installing pandas version: ${version}" | |
| # This loop tests cudf.pandas compatibility with older pandas-numpy versions, | |
| # requiring numpy<2. cupy>=14 dropped support for numpy<2, so we explicitly | |
| # downgrade cupy here to avoid an import failure when cupy tries | |
| # to load against the older numpy. | |
| set +e | |
| # We're iterating over a range of versions that might not all have | |
| # supported wheels, so our best bet for now is to disallow source | |
| # distributions. | |
| rapids-pip-retry install "numpy>=1.26,<2.0a0" "pandas==${version}" "cupy-cuda${RAPIDS_CUDA_VERSION%%.*}x<14" --only-binary=:all: | |
| INSTALL_SUCCESS=$? | |
| set -e | |
| if [[ ${INSTALL_SUCCESS} -ne 0 ]]; then | |
| echo "Failed to install pandas ${version} with numpy<2 for Python ${RAPIDS_PY_VERSION}. Skipping tests for this version." | |
| continue | |
| fi | |
| rapids-logger "Testing cudf.pandas compatibility with pandas ${version}.*" | |
| # Generate requirements for this pandas compat version. | |
| # Each entry pins numpy<2 + the specific pandas minor line + the CUDA-appropriate cupy<14. | |
| # cupy>=14 dropped support for numpy<2 (see https://github.com/cupy/cupy/issues/9709). | |
| pandas_requirements_txt="pandas-compat-${version}-requirements.txt" | |
| rapids-dependency-file-generator \ | |
| --config dependencies.yaml \ | |
| --file-key test_cudf_pandas_compat \ | |
| --output requirements \ | |
| --matrix "cuda=${RAPIDS_CUDA_VERSION};pandas_compat_version=${version}" \ | |
| > "${pandas_requirements_txt}" | |
| env_name="venv_pandas_${version}" | |
| python -m venv --clear "${env_name}" | |
| # shellcheck disable=SC1090 | |
| source "${env_name}/bin/activate" | |
| # notes: | |
| # | |
| # * echo to expand wildcard before adding `[test,cudf-pandas-tests]` requires for pip | |
| # * need to provide --constraint="${PIP_CONSTRAINT}" because that environment variable is | |
| # ignored if any other --constraint are passed via the CLI | |
| # | |
| rapids-pip-retry install \ | |
| -v \ | |
| --constraint ./constraints.txt \ | |
| --constraint "${PIP_CONSTRAINT}" \ | |
| "$(echo "${CUDF_WHEELHOUSE}"/cudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)[test,cudf-pandas-tests]" \ | |
| "$(echo "${LIBCUDF_WHEELHOUSE}"/libcudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)" \ | |
| "$(echo "${PYLIBCUDF_WHEELHOUSE}"/pylibcudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)" \ | |
| -r "${pandas_requirements_txt}" | |
| python -m pytest -p cudf.pandas \ | |
| if version_lte "${RAPIDS_PY_VERSION}" "3.13"; then | |
| if ! nvidia-smi -L >/dev/null 2>&1; then | |
| rapids-logger "GPU not detected; cannot run cudf.pandas compatibility tests" | |
| exit 1 | |
| fi | |
| for version in "${versions[@]}"; do | |
| rapids-logger "Testing cudf.pandas compatibility with pandas ${version}.*" | |
| # Generate requirements for this pandas compat version. | |
| # Each entry pins numpy<2 + the specific pandas minor line + the CUDA-appropriate cupy<14. | |
| # cupy>=14 dropped support for numpy<2 (see https://github.com/cupy/cupy/issues/9709). | |
| pandas_requirements_txt="pandas-compat-${version}-requirements.txt" | |
| rapids-dependency-file-generator \ | |
| --config dependencies.yaml \ | |
| --file-key test_cudf_pandas_compat \ | |
| --output requirements \ | |
| --matrix "cuda=${RAPIDS_CUDA_VERSION};pandas_compat_version=${version}" \ | |
| > "${pandas_requirements_txt}" | |
| env_name="venv_pandas_${version}" | |
| python -m venv --clear "${env_name}" | |
| # shellcheck disable=SC1090 | |
| source "${env_name}/bin/activate" | |
| # notes: | |
| # | |
| # * echo to expand wildcard before adding `[test,cudf-pandas-tests]` requires for pip | |
| # * need to provide --constraint="${PIP_CONSTRAINT}" because that environment variable is | |
| # ignored if any other --constraint are passed via the CLI | |
| # | |
| rapids-pip-retry install \ | |
| -v \ | |
| --constraint ./constraints.txt \ | |
| --constraint "${PIP_CONSTRAINT}" \ | |
| "$(echo "${CUDF_WHEELHOUSE}"/cudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)[test,cudf-pandas-tests]" \ | |
| "$(echo "${LIBCUDF_WHEELHOUSE}"/libcudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)" \ | |
| "$(echo "${PYLIBCUDF_WHEELHOUSE}"/pylibcudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)" \ | |
| -r "${pandas_requirements_txt}" | |
| python -m pytest -p cudf.pandas \ |
🧰 Tools
🪛 Shellcheck (0.11.0)
[info] 108-108: Not following: ./bin/activate was not specified as input (see shellcheck -x).
(SC1091)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@ci/cudf_pandas_scripts/run_tests.sh` around lines 90 - 125, The loop that
sets up env_name, activates the venv and installs deps assumes a GPU is present
but never verifies it; add a fail-fast GPU preflight check (e.g., run
`nvidia-smi -L` and exit non-zero with a clear error message) inside the loop
after sourcing "${env_name}/bin/activate" and before the `python -m pytest -p
cudf.pandas` invocation so tests abort early on CI when no GPU is available.
Comment on lines
+48
to
+69
| env_name="venv_polars_${version}" | ||
| python -m venv --clear "${env_name}" | ||
| # shellcheck disable=SC1090 | ||
| source "${env_name}/bin/activate" | ||
|
|
||
| rapids-logger "Installing cudf_polars and its dependencies for polars ${version}.*" | ||
|
|
||
| # notes: | ||
| # | ||
| # * echo to expand wildcard before adding `[test]` requires for pip | ||
| # * just providing --constraint="${PIP_CONSTRAINT}" to be explicit, and because | ||
| # that environment variable is ignored if any other --constraint are passed via the CLI | ||
| # | ||
| rapids-pip-retry install \ | ||
| -v \ | ||
| --prefer-binary \ | ||
| --constraint "${PIP_CONSTRAINT}" \ | ||
| "$(echo "${CUDF_POLARS_WHEELHOUSE}"/cudf_polars_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)[test,dask,ray]" \ | ||
| "$(echo "${LIBCUDF_WHEELHOUSE}"/libcudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)" \ | ||
| "$(echo "${PYLIBCUDF_WHEELHOUSE}"/pylibcudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)" \ | ||
| -r "polars-compat-${version}-requirements.txt" | ||
|
|
There was a problem hiding this comment.
Guard venv setup/install failures to avoid running tests in a bad environment.
With set +e, failures in venv creation/activation/dependency install do not stop execution. The script can continue into test execution with a partially configured or wrong environment.
Suggested patch
- python -m venv --clear "${env_name}"
+ if ! python -m venv --clear "${env_name}"; then
+ EXITCODE=1
+ FAILED+=("${version}")
+ rapids-logger "Failed to create venv for polars ${version}.*"
+ continue
+ fi
# shellcheck disable=SC1090
- source "${env_name}/bin/activate"
+ if ! source "${env_name}/bin/activate"; then
+ EXITCODE=1
+ FAILED+=("${version}")
+ rapids-logger "Failed to activate venv for polars ${version}.*"
+ rm -rf "${env_name}" "${polars_requirements_txt}"
+ continue
+ fi
...
- rapids-pip-retry install \
+ if ! rapids-pip-retry install \
-v \
--prefer-binary \
--constraint "${PIP_CONSTRAINT}" \
"$(echo "${CUDF_POLARS_WHEELHOUSE}"/cudf_polars_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)[test,dask,ray]" \
"$(echo "${LIBCUDF_WHEELHOUSE}"/libcudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)" \
"$(echo "${PYLIBCUDF_WHEELHOUSE}"/pylibcudf_"${RAPIDS_PY_CUDA_SUFFIX}"*.whl)" \
- -r "polars-compat-${version}-requirements.txt"
+ -r "polars-compat-${version}-requirements.txt"; then
+ EXITCODE=1
+ FAILED+=("${version}")
+ rapids-logger "Dependency install failed for polars ${version}.*"
+ deactivate
+ rm -rf "${env_name}" "${polars_requirements_txt}"
+ continue
+ fiAlso applies to: 91-103
🧰 Tools
🪛 Shellcheck (0.11.0)
[info] 51-51: Not following: ./bin/activate was not specified as input (see shellcheck -x).
(SC1091)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@ci/test_wheel_cudf_polars.sh` around lines 48 - 69, The venv
creation/activation and dependency installation steps (env_name, python -m venv
--clear, source "${env_name}/bin/activate", and rapids-pip-retry install) are
currently unguarded so failures are ignored; modify the script to immediately
stop and report an error if any of these steps fail—either enable errexit (set
-e) around the venv/install block or explicitly check the exit status after
python -m venv, after source "${env_name}/bin/activate", and after
rapids-pip-retry install and call a failure handler (logging via rapids-logger
and exit nonzero) so tests never run in a broken environment; apply the same
guarding pattern to the later venv block referenced for lines 91-103.
Comment on lines
+24
to
+30
| values = matrix.get(matrix_var) | ||
| if values is None: | ||
| raise RuntimeError( | ||
| f"Matrix variable '{matrix_var}' not found in file key '{file_key}'" | ||
| ) | ||
|
|
||
| return [str(v) for v in values] |
There was a problem hiding this comment.
Validate matrix value type before iterating.
On Line 30, a scalar YAML value would be iterated as characters and silently produce invalid output. Fail fast unless the matrix value is a sequence.
Suggested patch
- values = matrix.get(matrix_var)
+ values = matrix.get(matrix_var)
if values is None:
raise RuntimeError(
f"Matrix variable '{matrix_var}' not found in file key '{file_key}'"
)
+ if not isinstance(values, (list, tuple)):
+ raise RuntimeError(
+ f"Matrix variable '{matrix_var}' in file key '{file_key}' must be a list"
+ )
return [str(v) for v in values]🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@ci/utils/get_matrix_values.py` around lines 24 - 30, The code currently
treats the fetched `values` as iterable, which will iterate characters if a
scalar string is returned; update the block that reads `values =
matrix.get(matrix_var)` to validate the type before iterating: check that
`values` is a sequence (e.g., list/tuple or a collections.abc.Sequence) and
explicitly reject strings/bytes, and if it is not a proper sequence raise a
RuntimeError (include `matrix_var` and `file_key` in the message) instead of
silently iterating; then safely return `[str(v) for v in values]`.
vyasr
approved these changes
May 30, 2026
|
|
||
| # Generate requirements for this pandas compat version. | ||
| # Each entry pins numpy<2 + the specific pandas minor line + the CUDA-appropriate cupy<14. | ||
| # cupy>=14 dropped support for numpy<2 (see https://github.com/cupy/cupy/issues/9709). |
Contributor
There was a problem hiding this comment.
Note that this issue was resolved, so we only have to exclude cupy==14.0.0, later 14.x releases are fine for the CAI issue. However, numpy 2 support was indeed dropped, which I think is the main point, so maybe we just need a different link here?
Contributor
|
/merge |
This was referenced Jun 3, 2026
This was referenced Jun 22, 2026
This was referenced Jun 30, 2026
Merged
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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.
Description
Replaces the pip index versions PyPI network call in ci/cudf_pandas_scripts/run_tests.sh and
ci/test_wheel_cudf_polars.shwith explicit version lists defined in dependencies.yaml. Adds a newci/utils/get_matrix_values.pyutility to read those lists. For each version, a per-version requirements.txt is generated via rapids-dependency-file-generator and separate virtual environment is created with all dependency including downloaded wheels, avoiding environment contamination between runs.Checklist