Skip to content
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

Updates to build and test nx-cugraph wheel as part of CI and nightly workflows #3852

Merged

Conversation

rlratzel
Copy link
Contributor

@rlratzel rlratzel commented Sep 8, 2023

closes rapidsai/graph_dl#302

  • Updates GHA yaml files to build and test a nx-cugraph wheel
  • Adds CI scripts for building and testing the nx-cugraph wheel
  • Adds a smoketest script for nx-cugraph
  • Relevant code cleanup: removes unnecessary dataset download from cugraph wheel testing

…unnecessary dataset download from cugraph wheel testing.
@rlratzel rlratzel added improvement Improvement / enhancement to an existing function non-breaking Non-breaking change labels Sep 8, 2023
@rlratzel rlratzel added this to the 23.10 milestone Sep 8, 2023
@rlratzel rlratzel self-assigned this Sep 8, 2023
@copy-pr-bot

This comment was marked as resolved.

@rlratzel
Copy link
Contributor Author

rlratzel commented Sep 8, 2023

/ok to test

@rlratzel

This comment was marked as duplicate.

@rlratzel

This comment was marked as duplicate.

@rlratzel

This comment was marked as duplicate.

@rlratzel

This comment was marked as duplicate.

@rlratzel

This comment was marked as duplicate.

…he script, uses ls instead of echo to expand wildcards so errors for missing files are more obvious.
@rlratzel

This comment was marked as duplicate.

@rlratzel rlratzel marked this pull request as ready for review September 22, 2023 20:32
@rlratzel rlratzel requested a review from a team as a code owner September 22, 2023 20:32
@@ -9,9 +9,9 @@ package_dir=$2
mkdir -p ./dist
RAPIDS_PY_CUDA_SUFFIX="$(rapids-wheel-ctk-name-gen ${RAPIDS_CUDA_VERSION})"

# echo to expand wildcard before adding `[extra]` requires for pip
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There was a case where the wildcard did not match the .whl files on disk and 'echo' simply echo'd the wildcard pattern, which made pip generate a misleading error message. Using 'ls' will result in the script erroring out with a clear message about the missing/mis-named files.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Odd, but I'm fine with the change if it works.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree with this change, I see the problem with shell wildcard expansion.

@rlratzel rlratzel requested review from vyasr and bdice September 22, 2023 20:50
@rlratzel

This comment was marked as duplicate.

@rlratzel

This comment was marked as duplicate.

@rlratzel

This comment was marked as duplicate.

Copy link
Contributor

@vyasr vyasr left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Couple of notes, but looks fine otherwise.

ci/check_style.sh Outdated Show resolved Hide resolved
@@ -9,9 +9,9 @@ package_dir=$2
mkdir -p ./dist
RAPIDS_PY_CUDA_SUFFIX="$(rapids-wheel-ctk-name-gen ${RAPIDS_CUDA_VERSION})"

# echo to expand wildcard before adding `[extra]` requires for pip
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Odd, but I'm fine with the change if it works.

pushd ./datasets
bash ./get_test_data.sh
popd
fi
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I guess this change is necessary for nightlies? Were nightly tests for arm not running the tests that required datasets?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We shouldn't remove this, but might need to modify it to only be skipped in ARM PR jobs. In ARM PR jobs, we can only run "smoke tests" due to limited resource capacity.

# Run smoke tests for aarch64 pull requests

Downloading the datasets is time-intensive and should not be performed for "smoke tests" in ARM PR jobs because it wastes a lot of time on the ARM GPU node.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm terribly sorry. I read this backwards. This removed downloading datasets for ALL jobs (all arches, nightly/PR). Are we not dependent on datasets for tests to pass anymore?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The get_test_data.sh script will download supplemental datasets that are too big to commit to our github repo. Those are currently only used by C++ tests. The python tests use the smaller .csv datasets committed to the repo here, so it's safe to skip downloading when only testing python code.

I only removed it here since it was in the proximity of the changes for the wheel builds, but I think an audit of what runs require downloading the supplemental datasets would be good for a separate PR.

Prior to the recent commits to the PR which triggered CI running now, we had everything passing with the download step removed.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @rlratzel, I'll approve now.

@rlratzel

This comment was marked as duplicate.

Copy link
Contributor

@bdice bdice left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Requesting changes. We must skip dataset downloads for ARM smoke tests (ARM PRs).

@@ -9,9 +9,9 @@ package_dir=$2
mkdir -p ./dist
RAPIDS_PY_CUDA_SUFFIX="$(rapids-wheel-ctk-name-gen ${RAPIDS_CUDA_VERSION})"

# echo to expand wildcard before adding `[extra]` requires for pip
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree with this change, I see the problem with shell wildcard expansion.

pushd ./datasets
bash ./get_test_data.sh
popd
fi
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We shouldn't remove this, but might need to modify it to only be skipped in ARM PR jobs. In ARM PR jobs, we can only run "smoke tests" due to limited resource capacity.

# Run smoke tests for aarch64 pull requests

Downloading the datasets is time-intensive and should not be performed for "smoke tests" in ARM PR jobs because it wastes a lot of time on the ARM GPU node.

@rlratzel
Copy link
Contributor Author

/merge

@rapids-bot rapids-bot bot merged commit 5d2f548 into rapidsai:branch-23.10 Sep 27, 2023
67 checks passed
@raydouglass raydouglass mentioned this pull request Oct 2, 2023
rapids-bot bot pushed a commit that referenced this pull request Oct 17, 2023
A couple PRs were merged after `branch-23.12` was created and contained RAPIDS versions that need to be updated in `branch-23.12`.

Ref:
- #3838
- #3852

Authors:
  - Ray Douglass (https://github.com/raydouglass)
  - Rick Ratzel (https://github.com/rlratzel)

Approvers:
  - AJ Schmidt (https://github.com/ajschmidt8)
  - Brad Rees (https://github.com/BradReesWork)

URL: #3905
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
improvement Improvement / enhancement to an existing function non-breaking Non-breaking change
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants