Garmin Running Data Normalizer v1.4.0

Garmin Running Data Normalizer v1.4.0

These are the release notes for the v1.4.0 Export Evidence and Diagnostics
release. Version 1.4.0 is the current stable release, published from the
release-ready main source after the v1.4.0 tag, GitHub Release, and Production
PyPI publication gates.

Added

• Export Evidence Doctor for bounded pre-run readiness and completed-output
  authority checks.
• diagnostics/source_completeness.json for source-family observations.
• diagnostics/run_quality.json as a machine-readable projection of completed
  Run-All evidence.
• A typed, deterministic, public-safe six-member Support Bundle that requires
  Human review and never uploads automatically.
• First-user guidance for interpreting PASS, PASS_WITH_WARNINGS,
  PARTIAL_SUCCESS, and Fatal outcomes.

Compatibility

• The 17-dataset, 212-field, 6-relationship inventory is unchanged.
• Stable keys, normalized truth, Snapshot semantics, Run-All semantics, and the
  Activities-only External-safe Pack remain unchanged.
• The Product exit mapping remains PASS -> 0,
  PASS_WITH_WARNINGS -> 0, PARTIAL_SUCCESS -> 3, and Fatal -> 2.
• PARTIAL_SUCCESS / exit 3 is a completed result with bounded FIT exclusions,
  not a Fatal result.

Privacy and limitations

All processing remains local-first. The Support Bundle excludes raw rows,
identifiers, paths, coordinates, exact timestamps, private digests, environment
inventory, and free-form exception text. It is a diagnostic derivative, not the
existing Activities-only External-safe Analysis Pack, and it must be reviewed
before sharing. The release makes no universal Garmin, medical, or coaching
claim.

Install

python -m pip install garmin-running-data-normalizer==1.4.0
garmin-running-data-normalizer --version