Update Broken-links-checker-final.yml

Author: Vamshi-Microsoft

.github/workflows/Broken-links-checker-final.yml

@@ -1,4 +1,4 @@
-name: Broken Link Checker
+name: Smart Broken Link Checker
 
 on:
   pull_request:
@@ -10,51 +10,93 @@ permissions:
   contents: read
 
 jobs:
-  markdown-link-check:
-    name: Check Markdown Broken Links
+  smart-link-check:
+    name: Filtered Markdown Broken Links
     runs-on: ubuntu-latest
 
     steps:
-      - name: Checkout Repo
+      - name: Checkout Repository
         uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
-      - name: Get Added/Modified Markdown Files (PR only)
-        id: changed-files
-        if: github.event_name == 'pull_request'
+      - name: Install lychee
         run: |
-          git fetch origin ${{ github.base_ref }}
-          files=$(git diff --name-only origin/${{ github.base_ref }}...HEAD | grep '\.md$' || true)
-          echo "md_files<<EOF" >> $GITHUB_OUTPUT
-          echo "$files" >> $GITHUB_OUTPUT
-          echo "EOF" >> $GITHUB_OUTPUT
-      - name: Check Broken Links in Added/Modified Files (PR)
-        if: github.event_name == 'pull_request' && steps.changed-files.outputs.md_files != ''
-        uses: lycheeverse/[email protected]
-        with:
-          args: >
-            --verbose --exclude-mail --no-progress --exclude ^https?://
-            ${{ steps.changed-files.outputs.md_files }}
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      
-      - name: Preprocess All Markdown Files (strip comments)
-        if: github.event_name == 'workflow_dispatch'
+          curl -Ls https://github.com/lycheeverse/lychee/releases/latest/download/lychee-linux.zip -o lychee.zip
+          unzip lychee.zip -d lychee-bin
+          chmod +x lychee-bin/lychee
+          echo "$PWD/lychee-bin" >> $GITHUB_PATH
+
+      - name: Determine Target Files
+        id: target-files
+        run: |
+          if [[ "${{ github.event_name }}" == "pull_request" ]]; then
+            git fetch origin ${{ github.base_ref }}
+            files=$(git diff --name-only origin/${{ github.base_ref }}...HEAD | grep '\.md$' || true)
+          else
+            files=$(find . -type f -name "*.md")
+          fi
+          echo "$files" > all-md-files.txt
+          echo "files=$(cat all-md-files.txt | paste -sd ' ' -)" >> $GITHUB_OUTPUT
+
+      - name: Run Lychee (Initial)
+        run: |
+          mkdir -p lychee-output
+          lychee --verbose --no-progress --exclude-mail --format json \
+            ${{ steps.target-files.outputs.files }} > lychee-output/report.json
+          # Also produce plain-text report for human reading
+          lychee --verbose --no-progress --exclude-mail \
+            ${{ steps.target-files.outputs.files }} > lychee-output/report.txt || true
+
+      - name: Extract Failed Files
+        id: extract-failed
+        run: |
+          mkdir -p sanitize-targets
+          jq -r '.failures[].input' lychee-output/report.json | sort -u > failed-files.txt
+          if [[ -s failed-files.txt ]]; then
+            echo "files_to_sanitize=$(cat failed-files.txt | paste -sd ' ' -)" >> $GITHUB_OUTPUT
+          fi
+
+      - name: Sanitize Failed Files (Remove Comments)
+        if: steps.extract-failed.outputs.files_to_sanitize != ''
         run: |
-          mkdir -p cleaned_md
-          find . -name "*.md" ! -path "./cleaned_md/*" | while read file; do
-            mkdir -p "cleaned_md/$(dirname "$file")"
-            sed -E '/<!--/,/-->/d' "$file" > "cleaned_md/$file"
-          done
-      
-      - name: Check Broken Links in Entire Repo (Manual)
-        if: github.event_name == 'workflow_dispatch'
-        uses: lycheeverse/[email protected]
+          mkdir -p sanitized_md
+          while IFS= read -r file; do
+            [ -f "$file" ] || continue
+            mkdir -p "sanitized_md/$(dirname "$file")"
+            perl -0777 -pe 's/<!--.*?-->//gs' "$file" > "sanitized_md/$file"
+          done < failed-files.txt
+
+      - name: Re-run Lychee on Sanitized Files
+        if: steps.extract-failed.outputs.files_to_sanitize != ''
+        run: |
+          mkdir -p lychee-output-sanitized
+          while IFS= read -r file; do
+            [ -f "sanitized_md/$file" ] || continue
+            echo "🔍 Rechecking: sanitized_md/$file"
+            lychee --verbose --no-progress --exclude-mail "sanitized_md/$file" || echo "$file" >> final-failures.txt
+          done < failed-files.txt
+
+      - name: Upload Lychee Reports
+        uses: actions/upload-artifact@v4
         with:
-          args: >
-            --verbose --exclude-mail --no-progress --exclude ^https?://
-            cleaned_md/**/*.md
-          output: lychee/out.md
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          name: lychee-original-reports
+          path: lychee-output
+
+      - name: Upload Sanitized Report Failures
+        if: steps.extract-failed.outputs.files_to_sanitize != ''
+        uses: actions/upload-artifact@v4
+        with:
+          name: lychee-sanitized-failures
+          path: |
+            final-failures.txt
+            sanitized_md
+
+      - name: Fail if Broken Links Remain
+        if: failure() && steps.extract-failed.outputs.files_to_sanitize != ''
+        run: |
+          if [ -s final-failures.txt ]; then
+            echo "🚫 Broken links still exist after comment-stripping:"
+            cat final-failures.txt
+            exit 1
+          fi