github-actions

GitHub Actions quality gate

Use DataSentry to stop a pull request or release when a dataset contains high-severity quality issues, while still uploading a human-readable report for review.

Copy datasentry-quality-gate.yml to .github/workflows/data-quality.yml and change the dataset path:

- name: Scan data and enforce quality gate
  run: datasentry scan data/orders.csv --fail-on high

The example deliberately keeps repair out of CI. Automated pipelines should detect, explain, and gate; state-changing repairs should stay reviewable and reversible.

What the workflow does

1. installs Python and datasentry-ai;
2. scans the selected dataset with deterministic detectors;
3. exits non-zero if the configured severity gate fails;
4. exports the latest scan as an HTML report even when the gate fails;
5. uploads the report as a workflow artifact.

Other report formats

DataSentry can also export JSON, Markdown, JUnit, and SARIF:

datasentry report export latest --as junit --output datasentry-junit.xml
datasentry report export latest --as sarif --output datasentry.sarif.json

Use JUnit when your CI surface understands test reports and SARIF when you want to feed findings into compatible code-scanning tooling.