parse.yml

name: parse

# Daily freshness sweep + per-crate parse jobs.
#
# All work happens in the `codeview` Rust CLI now — TypeScript scripts
# are gone.  Each job calls one `codeview cron *` subcommand:
#
#   freshness → `codeview cron sweep`
#   parse     → `codeview cron parse-one`
#   catalog   → `codeview cron catalog`
#
# Std crates are out of scope here — see std.yml.

on:
  schedule:
    - cron: '0 3 * * *'
  workflow_dispatch:
    inputs:
      force:
        description: 'Comma-separated crate names to force-parse (skip freshness)'
        required: false
        default: ''
      max_crates:
        description: 'Cap on stale crates parsed this run'
        required: false
        default: '20'
      watchlist:
        description: "'catalog' | 'top:N' | path/to/file"
        required: false
        default: 'catalog'

env:
  R2_BUCKET: crate-graphs
  STATIC_R2_TARGET: remote
  PARSER_REVISION: ${{ github.sha }}
  # R2 over the S3-compatible endpoint
  R2_ACCESS_KEY_ID: ${{ secrets.R2_ACCESS_KEY_ID }}
  R2_SECRET_ACCESS_KEY: ${{ secrets.R2_SECRET_ACCESS_KEY }}
  R2_ACCOUNT_ID: ${{ secrets.R2_ACCOUNT_ID }}
  # Skip the SvelteKit bundle that codeview-cli's build.rs embeds — the
  # cron subcommands don't need it.
  CODEVIEW_SKIP_SIDECAR: '1'

jobs:
  freshness:
    name: freshness sweep
    runs-on: ubuntu-latest
    outputs:
      matrix: ${{ steps.sweep.outputs.matrix }}
      count: ${{ steps.sweep.outputs.count }}
    steps:
      - uses: actions/checkout@v6

      - name: Install Rust stable
        uses: dtolnay/rust-toolchain@3c5f7ea28cd621ae0bf5283f0e981fb97b8a7af9
        with:
          toolchain: stable

      - uses: Swatinem/rust-cache@v2
        with:
          shared-key: cron-release

      - name: Build codeview-cli (release)
        run: cargo build --release -p codeview-cli

      - name: Run sweep
        id: sweep
        run: |
          target/release/codeview cron sweep \
            --watchlist "${{ inputs.watchlist || 'catalog' }}" \
            --max-crates "${{ inputs.max_crates || '20' }}" \
            --force "${{ inputs.force || '' }}" \
            --bucket "$R2_BUCKET" \
            --quiet

  parse:
    name: parse ${{ matrix.crate.name }}@${{ matrix.crate.version }}
    needs: freshness
    if: needs.freshness.outputs.count != '0' && needs.freshness.outputs.count != ''
    runs-on: ubuntu-latest
    strategy:
      fail-fast: false
      max-parallel: 10
      matrix: ${{ fromJSON(needs.freshness.outputs.matrix) }}
    steps:
      - uses: actions/checkout@v6

      - name: Install Rust nightly (for the codeview-rustdoc parser path)
        uses: dtolnay/rust-toolchain@3c5f7ea28cd621ae0bf5283f0e981fb97b8a7af9
        with:
          toolchain: nightly

      - uses: Swatinem/rust-cache@v2
        with:
          shared-key: cron-release

      - name: Build codeview-cli (release)
        run: cargo build --release -p codeview-cli

      - name: Parse + publish + record freshness
        env:
          NAME: ${{ matrix.crate.name }}
          VERSION: ${{ matrix.crate.version }}
        # codeview cron parse-one exit codes:
        #   0  ok
        #   64 transient (retry next sweep)
        #   65 permanent (don't retry until parser/schema changes)
        #   70 internal bug — page a human
        run: |
          set +e
          target/release/codeview cron parse-one \
            --name "$NAME" \
            --version "$VERSION" \
            --bucket "$R2_BUCKET"
          code=$?
          set -e
          case $code in
            0)  echo "::notice title=parse $NAME@$VERSION::ok" ;;
            64) echo "::warning title=parse $NAME@$VERSION::transient — will retry next sweep" ;;
            65) echo "::warning title=parse $NAME@$VERSION::permanent — skipping until parser/schema changes" ;;
            *)  echo "::error title=parse $NAME@$VERSION::internal error (exit $code)" ;;
          esac
          # Translate handled non-zero codes back to 0 so they don't flag
          # the whole job as failed.
          exit $([ $code -le 65 ] && echo 0 || echo $code)

  catalog:
    name: rebuild catalog
    needs: [freshness, parse]
    if: always() && needs.freshness.outputs.count != '0' && needs.freshness.outputs.count != ''
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v6

      - name: Install Rust stable
        uses: dtolnay/rust-toolchain@3c5f7ea28cd621ae0bf5283f0e981fb97b8a7af9
        with:
          toolchain: stable

      - uses: Swatinem/rust-cache@v2
        with:
          shared-key: cron-release

      - name: Build codeview-cli (release)
        run: cargo build --release -p codeview-cli

      - name: Rebuild rust/catalog.json from freshness index
        run: target/release/codeview cron catalog --bucket "$R2_BUCKET"

  summary:
    name: parse summary
    needs: [freshness, parse, catalog]
    if: always() && needs.freshness.outputs.count != '0'
    runs-on: ubuntu-latest
    steps:
      - name: Echo summary
        run: |
          echo "freshness produced ${{ needs.freshness.outputs.count }} stale crates"
          echo "parse job result: ${{ needs.parse.result }}"
          echo "catalog job result: ${{ needs.catalog.result }}"