Skip to content

report-artifacts.md

Latest commit

 

History

History
130 lines (89 loc) · 3.9 KB

report-artifacts.md

File metadata and controls

130 lines (89 loc) · 3.9 KB
Raw

Back to documentation index

Report artifacts

This reference describes the files produced under artifacts/latest/site/.

Generated files

  • index.html
  • report.html
  • run.json
  • snippets.json
  • openfoodfacts-data-quality-json.zip

The JSON files are the structured artifacts. The ZIP file bundles those JSON outputs for download.

index.html and report.html render the same content.

HTML report

The HTML report is the review summary for one migration run.

It lists compared checks that match. It also lists compared checks with missing or extra findings, runtime only checks, and product codes in retained mismatch examples.

The page header always includes the run id and source snapshot id.

The report is a summary view. It does not inline every finding from the run.

When the source snapshot contains rows that were skipped before execution, the report also shows a source input warning with the skipped row count and example locations.

Summary counters

  • Mismatching: Checks with at least one missing or extra finding under strict comparison
  • Missing Findings: Findings present on the reference side and absent on the migrated side
  • Extra Findings: Findings present on the migrated side and absent on the reference side
  • Affected Products: Unique product codes that appear in retained mismatch examples
  • Runtime Only: Active checks with no legacy comparison baseline

Check cards

Each check card includes:

  • the canonical check id
  • the definition language, python or dsl
  • the run outcome bucket
  • matched, missing, and extra counts when comparison applies
  • retained mismatch examples for each side
  • implementation and legacy snippets when snippet provenance is available

If legacy source provenance is unavailable, the card still renders and explains that state in the snippet area.

Retained mismatch examples are capped by MISMATCH_EXAMPLES_LIMIT.

run.json

run.json is the canonical structured output of the migration run model.

Use run.json for:

  • process run results after execution
  • build another viewer or dashboard
  • diff runs programmatically
  • archive one run in a stable structured format

The root payload includes kind and schema_version metadata before the run body.

run.json contains exact run totals plus the retained mismatch examples that fit inside the configured cap. It omits UI fields used only by the report:

  • run_outcome
  • snippet panel payloads
  • snippet warning messages

The root payload also includes source_input, which records source boundary diagnostics such as skipped-row counts and example locations.

snippets.json

snippets.json stores structured code excerpts keyed by check id.

It includes:

  • implementation snippets from this repository
  • legacy source snippets when they can be resolved

The snippet origin values are implementation and legacy.

Each check entry also includes legacy_snippet_status with one of these values:

  • available
  • not_applicable
  • unavailable

Use snippets.json when you want provenance and review context without parsing HTML.

Scope

The renderer supports compared checks and checks that run without comparison in the same run.

Strict comparison counts and mismatch examples apply only to checks with a legacy baseline. Checks that run without comparison still contribute to run composition and per check output.

See also

Back to documentation index