Update docs-quality.yml

Author: Amanda-dong

.github/workflows/docs-quality.yml

@@ -1,118 +1,62 @@
 name: Docs Quality
+
 on:
   pull_request:
-  workflow_dispatch:
+    paths:
+      - "docs/**/*.md"
+      - "docs/**/*.mdx"
+      - "_typos.toml"
+      - ".github/workflows/docs-quality.yml"
+  push:
+    branches: [main]
+  workflow_dispatch: {}
+
+permissions:
+  contents: read
+
 jobs:
   typos:
     name: Spelling (typos-cli)
     runs-on: ubuntu-latest
     steps:
-      - name: Checkout
-        uses: actions/checkout@v6
+      - uses: actions/checkout@v4
       - uses: crate-ci/typos@master
 
   harper:
-    name: Grammar and Style Check
+    name: Grammar & headings (Harper)
     runs-on: ubuntu-latest
     continue-on-error: true
     steps:
-      - name: Checkout
-        uses: actions/checkout@v6
-      - name: Restore Harper CLI cache
+      - uses: actions/checkout@v4
+
+      - name: Cache harper-cli binary
         id: cache-harper
-        uses: actions/cache/restore@v4
+        uses: actions/cache@v4
         with:
-          path: |
-            ~/.cargo/bin/harper-cli
-            ~/.cargo/registry
-            ~/.cargo/git
-          key: ${{ runner.os }}-harper-cli
+          path: ~/.cargo/bin/harper-cli
+          key: harper-cli-${{ runner.os }}-v1
           restore-keys: |
-            ${{ runner.os }}-harper-cli
+            harper-cli-${{ runner.os }}-
       - name: Report cache status
         run: |
           echo "cache-hit output: '${{ steps.cache-harper.outputs.cache-hit }}'"
-          ls -la ~/.cargo/bin/ 2>&1 || echo "~/.cargo/bin does not exist"
-          if [ -x ~/.cargo/bin/harper-cli ]; then
-            echo "RESULT: harper-cli binary present and executable"
-            ~/.cargo/bin/harper-cli --version || true
+          if [ -f ~/.cargo/bin/harper-cli ]; then
+            echo "harper-cli binary present at ~/.cargo/bin/harper-cli"
           else
-            echo "RESULT: harper-cli binary NOT present -- full reinstall will run"
+            echo "harper-cli binary NOT present -- will build from source"
           fi
-      - name: Install Harper CLI
+      - name: Install harper-cli
         if: steps.cache-harper.outputs.cache-hit != 'true'
         run: cargo install --locked --git https://github.com/Automattic/harper.git harper-cli
-      - name: Lint docs with Harper
-        run: |
-          set -uo pipefail
-          ONLY_RULES=(
-            UseTitleCase
-            RepeatedWords
-          )
-          ONLY_ARG="--only $(IFS=,; echo "${ONLY_RULES[*]}")"
-          out="$(mktemp)"
-          total=0
-          bad=0
-          while IFS= read -r -d '' file; do
-            total=$((total+1))
-            echo "::group::$file"
-            if harper_out="$(harper-cli lint $ONLY_ARG "$file" 2>&1)"; then
-              echo "no issues"
-            else
-              bad=$((bad+1))
-              echo "$harper_out"
-              echo "$harper_out" >> "$out"
-            fi
-            echo "::endgroup::"
-          done < <(
-            find docs src/pages \
-              -type f \( -name "*.md" -o -name "*.mdx" \) \
-              -print0 2>/dev/null
-          )
-          {
-            echo "### Harper summary"
-            echo "- Files scanned: $total"
-            echo "- Files with issues: $bad"
-            echo ""
-            echo "### Violations by rule"
-            echo '```'
-            grep -oE '\[[A-Za-z]+::[A-Za-z]+\]' "$out" | sort | uniq -c | sort -rn || true
-            echo '```'
-          } >> "$GITHUB_STEP_SUMMARY"
 
-          echo "== Harper: $bad/$total files with issues =="
-          [ "$bad" -eq 0 ]
-
-      - name: Annotate PR with Harper warnings
-        if: always()
+      - name: Lint docs with Harper
         run: |
           set -uo pipefail
-          ONLY_RULES=(
-            UseTitleCase
-            RepeatedWords
-          )
-          ONLY_ARG="--only $(IFS=,; echo "${ONLY_RULES[*]}")"
-          while IFS= read -r -d '' file; do
-            if json_out="$(harper-cli lint $ONLY_ARG --format json "$file" 2>/dev/null)"; then
-              :
-            else
-              printf '%s' "$json_out" | python3 -c "import json,sys;d=json.load(sys.stdin);esc=lambda s: s.replace('%','%25').replace('\r','%0D').replace('\n','%0A');[print('::warning file='+esc(f.get('file',''))+',line='+str(l.get('line',1))+',col='+str(l.get('column',1))+',title='+esc(l.get('rule','?'))+'::'+esc(l.get('message',''))) for f in d for l in f.get('lints',[])]" || true
-            fi
-          done < <(
-            find docs src/pages \
-              -type f \( -name "*.md" -o -name "*.mdx" \) \
-              -print0 2>/dev/null
-          )
-
-      - name: Save Harper CLI cache
-        if: always() && steps.cache-harper.outputs.cache-hit != 'true'
-        uses: actions/cache/save@v4
-        with:
-          path: |
-            ~/.cargo/bin/harper-cli
-            ~/.cargo/registry
-            ~/.cargo/git
-          key: ${{ runner.os }}-harper-cli            CompaniesProductsAndTrademarks
+          IGNORED_RULES=(
+            SpellCheck
+            OrthographicConsistency
+            ProperNouns
+            CompaniesProductsAndTrademarks
             GoogleNames MicrosoftNames AmazonNames AzureNames MetaNames
             LongSentences Hedging FillerWords
             ExpandStandardInputAndOutput ExpandTimeShorthands ExpandMemoryShorthands
@@ -122,7 +66,6 @@ jobs:
           )
           IGNORE_ARGS=()
           for r in "${IGNORED_RULES[@]}"; do IGNORE_ARGS+=(--ignore "$r"); done
-
           out="$(mktemp)"
           total=0
           bad=0
@@ -136,7 +79,6 @@ jobs:
             fi
             echo "::endgroup::"
           done < <(find docs \( -name '*.md' -o -name '*.mdx' \) -print0)
-
           titlecase=$(grep -c 'Capitalization::UseTitleCase' "$out" || true)
           {
             echo "### Harper summary"
@@ -150,89 +92,4 @@ jobs:
             echo '```'
           } >> "$GITHUB_STEP_SUMMARY"
           echo "== Harper: $bad/$total files with issues; $titlecase UseTitleCase violations =="
-          [ "$bad" -eq 0 ]            else
-              bad=$((bad+1))
-            fi
-            echo "::endgroup::"
-          done < <(
-            find docs src/pages \
-              -type f \( -name "*.md" -o -name "*.mdx" \) \
-              -print0 2>/dev/null
-          )
-          titlecase=$(grep -c 'Capitalization::UseTitleCase' "$out" || true)
-          {
-            echo "### Harper summary"
-            echo "- Files scanned: $total"
-            echo "- Files with issues: $bad"
-            echo "- Heading title case violations: $titlecase"
-            echo ""
-            echo "### Violations by rule"
-            echo '```'
-            grep -oE '\[[A-Za-z]+::[A-Za-z]+\]' "$out" | sort | uniq -c | sort -rn || true
-            echo '```'
-          } >> "$GITHUB_STEP_SUMMARY"
-          echo "== Harper: $bad/$total files with issues; $titlecase UseTitleCase violations =="
-          [ "$bad" -eq 0 ]            ExpandMemoryShorthands
-            ExpandDirectory
-            AnA
-            NoFrenchSpaces
-            InflectedVerbAfterTo
-            CapitalizePersonalPronouns
-            Dashes
-          )
-          IGNORE_ARGS=()
-          for rule in "${IGNORED_RULES[@]}"; do
-            IGNORE_ARGS+=(--ignore "$rule")
-          done
-          out="$(mktemp)"
-          total=0
-          bad=0
-          while IFS= read -r -d '' file; do
-            total=$((total+1))
-            echo "::group::$file"
-            if harper-cli lint "${IGNORE_ARGS[@]}" "$file" 2>&1 | tee -a "$out"; then
-              :
-            else
-              bad=$((bad+1))
-            fi
-            echo "::endgroup::"
-          done < <(
-            find docs src/pages \
-              -type f \( -name "*.md" -o -name "*.mdx" \) \
-              -print0 2>/dev/null
-          )
-          titlecase=$(grep -c 'Capitalization::UseTitleCase' "$out" || true)
-          {
-            echo "### Harper summary"
-            echo "- Files scanned: $total"
-            echo "- Files with issues: $bad"
-            echo "- Heading title case violations: $titlecase"
-            echo ""
-            echo "### Violations by rule"
-            echo '```'
-            grep -oE '\[[A-Za-z]+::[A-Za-z]+\]' "$out" | sort | uniq -c | sort -rn || true
-            echo '```'
-          } >> "$GITHUB_STEP_SUMMARY"
-          echo "== Harper: $bad/$total files with issues; $titlecase UseTitleCase violations =="
           [ "$bad" -eq 0 ]
-      
-      - name: Annotate PR with Harper warnings
-        if: always()
-        run: |
-          set -uo pipefail
-          ONLY_RULES=(
-            UseTitleCase
-            RepeatedWords
-          )
-          ONLY_ARG="--only $(IFS=,; echo "${ONLY_RULES[*]}")"
-          while IFS= read -r -d '' file; do
-            if json_out="$(harper-cli lint $ONLY_ARG --format json "$file" 2>/dev/null)"; then
-              :
-            else
-              printf '%s' "$json_out" | python3 -c "import json,sys;d=json.load(sys.stdin);esc=lambda s: s.replace('%','%25').replace('\r','%0D').replace('\n','%0A');[print('::warning file='+esc(f.get('file',''))+',line='+str(l.get('line',1))+',col='+str(l.get('column',1))+',title='+esc(l.get('rule','?'))+'::'+esc(l.get('message',''))) for f in d for l in f.get('lints',[])]" || true
-            fi
-          done < <(
-            find docs src/pages \
-              -type f \( -name "*.md" -o -name "*.mdx" \) \
-              -print0 2>/dev/null
-          )