codecov.yml

# Codecov configuration — https://docs.codecov.com/docs/codecov-yaml
coverage:
  status:
    project:
      # Fixed floors, not `auto`. `auto` compares against the base commit, and
      # the frontend report was re-baselined on 2026-09-02 (app/vite.config.ts
      # now measures the whole SPA source instead of only the modules a test
      # imported: 43 files became 201). That is a one-time step DOWN in the
      # reported total — honest, but every PR opened across it would have died
      # on the 1 % threshold. Measured right after the re-baseline: backend
      # 87.9 %, frontend 19.2 %, combined 59.2 %.
      default:
        target: 55%
        threshold: 1%
      backend:
        # The backend is genuinely tested; this floor is a real gate.
        target: 85%
        flags:
          - backend
      frontend:
        # Under the measured 19.2 %, deliberately: the gap IS the missing
        # component tests, and a floor that reds a PR for adding untested UI
        # would only push people back toward measuring less. Raise it as the
        # component tests land — that is the point of stating it as a number.
        target: 15%
        flags:
          - frontend
    patch:
      backend:
        target: 80%
        flags:
          - backend
      frontend:
        # Informational to start with — the Vitest suite is young; tighten
        # once the pure modules (pdf/lineatur/glyphs) carry tests.
        target: auto
        threshold: 5%
        flags:
          - frontend

# Only report coverage for the tested source (core/ + api/ + app/src).
# Top-level key per the Codecov schema.
ignore:
  - "alembic/"
  - "tests/"
  - "data/"
  - "docs/"
  - "tools/"

flags:
  backend:
    paths:
      - core/
      - api/
    carryforward: true
  frontend:
    paths:
      - app/src/
    carryforward: true

comment:
  layout: "reach,diff,flags,components"
  behavior: default
  require_changes: false