dbt Doctordbt DoctorStar on GitHub

Score Modes

Project score is always 0-100. The score is computed from three penalty buckets:

  • Rule variety penalty (unique rules hit)
  • File spread penalty (only in files mode)
  • Finding volume penalty (total diagnostic count)

Formula

score = max(0, round(100 - rulePenalty - filesPenalty - findingsPenalty))

Where:

  • rulePenalty = (unique_error_rules 1.5) + (unique_warning_rules 0.75)
  • filesPenalty = round((affectedFiles / totalFilesScanned) * 40) in files mode, else 0
  • findingsPenalty = min(15, floor(totalFindings / 7))

Labels

  • Great for score >= 75
  • Needs work for score >= 50 and < 75
  • Critical for score < 50

Modes

  • Files mode: default, penalizes how broadly issues are spread across files.
  • Unique-rules mode: ignores file spread, focuses on rule diversity + finding volume.

Configuration

In .dbt-doctor:

score_mode=files

or:

score_mode=unique-rules

CLI override:

npx dbt-doctor@latest --score-mode files