Skip to content

ALL-DEPE0004 · dependency_evidence_gap_percentage

Measure dependencies missing facts required by offline version checks.

This is a deterministic rule for all languages. Read its implementation.

Collect current dependency evidence in memory and divide records missing an exact resolved release date, latest compatible version, or latest compatible release date by all records. The selected provider, not a bundled package catalog, defines the target project’s evidence.

Every finding identifies the dependency, resolved version, missing fields, artifact source location, and any bounded collection failures on the record. The percentage is zero for an empty dependency set because there is no missing package evidence to measure. The value is the percentage of dependency records missing a required fact.

Local, VCS, private-index, and ambiguous environment resolutions can remain unknown, but their missing facts stay visible rather than being guessed. Projects may ignore this rule when an internal evidence provider owns those packages. This rule does not judge package quality, capability fit, or maintenance.

Two incomplete records among ten dependencies return 20. Complete evidence returns 0. A network failure records the missing fields and failure source in the current check.

  • Cites “Python Packaging User Guide”, Simple Repository API. Open reference
  • Cites “PyPI API documentation”, exact release metadata. Open reference
  • Cites “Python Packaging User Guide”, lock file specification. Open reference