Repository navigation
Conversation
PR checks picked components only from producer/<p>/ and consumer/ paths, so changes to the validator, report code, producer workflows or shared actions merged without the affected component ever running. - scripts/get_pr_components.py maps changed files to components: a component's files, its workflow or a shared action/script it uses run it for real; validation code changes run dbt and scenarios for real and re-validate the stored events of the last release run for spark_dataproc and hive_dataproc (no clusters). - The release check stores each validated Spark/Hive pair's events under gs://open-lineage-e2e/baseline-events/ (update-baseline input, off for every other caller). - revalidate_stored_events.yml validates that baseline with the PR's code, one report artifact per component, compared like any other report. - The spec fetch moves from the run_event_validation action into dev/validation/get_ol_specs.sh, shared with the re-validation. - Adds pytest setup and a script-tests job. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: tnazarew <tomasz.nazarewicz@getindata.com>
tnazarew
force-pushed
the
ci/pr-coverage
branch
from
November 12, 2026 13:16
0445bd7 to
8f79e0b
Compare
This was referenced Nov 16, 2026
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
PR checks decide what to run only from
producer/<p>/andconsumer/…paths. A change to the validator (scripts/validate_ol_events.py,report.py,compare_events.py), a producer workflow (.github/workflows/producer_*.yml) or a shared action (dataproc_cluster,run_event_validation) merges without the affected component ever running in CI.What
scripts/get_pr_components.py(tested) replaces the inline path check inmain_pr.yml:dataproc_cluster,get_openlineage_artifacts,upload_artifactsactions,get_valid_test_scenarios.shrun_event_validation, collect workflow)report.py/compare_events.py; re-validation of stored events for spark_dataproc and hive_dataproc (no clusters)update-baseline: trueto the Spark and Hive workflows, which then store each validated pair's events (and Spark'sstatus.json) undergs://open-lineage-e2e/baseline-events/<component>/<cv>/<ol>/. Every other caller leaves it off. The bucket's 90-day delete rule only coversevents/andlogs/.revalidate_stored_events.yml: one job per component validates every baseline pair with the PR's code and uploads one report artifact;collect-and-compare-reportscompares it withreport.jsonlike any other report, so new failures fail the PR. No baseline → a warning ("not covered"), not a failure.run_event_validationaction intodev/validation/get_ol_specs.sh(same behaviour), shared with the re-validation.scripts/tests/,scripts/requirements-dev.txt) and ascript-testsjob.Verification
pytest scripts/tests→ 25 passed (path map cases, incl. the old paths keeping their behaviour and anchored prefixes).get_ol_specs.sh, both pairs validated, one merged report keyed like the release entries.After merge
The baseline is empty until the next release run that runs Spark/Hive. I'll seed it once (with approval) by copying the last release run's stored Spark events and the latest Hive events into
baseline-events/, so the next validator PR is covered right away.Note: Hive currently fails its semantic checks, but no Hive report ever reaches the comparison (an upload path bug), so that isn't visible yet. A follow-up PR fixes the upload and the stale Hive expectations.
🤖 Generated with Claude Code