Adding link checker

Author: aireilly

.github/workflows/check-links.yml

@@ -0,0 +1,33 @@
+name: Check links in AsciiDoc
+
+on: [push, pull_request]
+
+jobs:
+  check-links:
+    name: Check links in modified files
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Install Asciidoctor
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y asciidoctor
+
+      - name: Make scripts executable
+        run: |
+          chmod +x scripts/check-links.sh
+          chmod +x scripts/check-modified.sh
+
+      - name: Fetch base branch
+        run: git fetch origin main
+
+      - name: Check links in modified files
+        shell: bash
+        run: |
+          scripts/check-modified.sh
+
+

scripts/check-links.sh

@@ -0,0 +1,77 @@
+#!/bin/bash
+#
+# Checks for 404 links using Asciidoctor and curl
+
+usage() {
+    echo "Usage: $0 [<adoc_file>]"
+    exit 1
+}
+
+# Parse arguments
+INPUT_FILE=""
+
+# Check dependencies
+if ! asciidoctor -v >/dev/null 2>&1; then
+    echo "Error: Asciidoctor is not installed" >&2
+    exit 1
+fi
+
+INPUT_FILE="$1"
+
+if [ $# -eq 0 ]; then
+    usage
+fi
+
+# Create temp file for flagging broken links
+TMP_FILE=$(mktemp)
+echo "0" > "$TMP_FILE"
+
+# Load ignore patterns from external file
+IGNORE_FILE="$(dirname "$0")/links.ignore"
+
+if [ ! -f "$IGNORE_FILE" ]; then
+    echo "Error: Missing ignore patterns file: $IGNORE_FILE" >&2
+    exit 1
+fi
+
+mapfile -t IGNORE_PATTERNS < "$IGNORE_FILE"
+PATTERNS_DECL=$(declare -p IGNORE_PATTERNS)
+
+check_url() {
+    local URL=$1
+    eval "$PATTERNS_DECL"
+
+    URL=${URL%[.,;:?!\]\)]}
+
+    for PATTERN in "${IGNORE_PATTERNS[@]}"; do
+        if [[ "$URL" =~ $PATTERN ]]; then
+            exit 0
+        fi
+    done
+
+    STATUS=$(curl -Ls -o /dev/null -w "%{http_code}" --max-time 5 --connect-timeout 2 "$URL")
+
+    if [[ "$STATUS" != "000" && "$STATUS" != "403" && ! "$STATUS" =~ ^(2|3)[0-9]{2}$ ]]; then
+        echo -e "Invalid URL (HTTP status $STATUS): \n\033[31m$URL\033[0m"
+        echo "1" > "$TMP_FILE"
+    fi
+}
+
+export TMP_FILE
+export -f check_url
+
+run_url_checks() {
+    local FILE="$1"
+    echo -e "\033[32mChecking: $FILE\033[0m"
+    asciidoctor "$FILE" -a doctype=book -o - | \
+        grep -Eo '(http|https)://[a-zA-Z0-9./?=%_-]*' | \
+        sort -u | \
+        xargs -P 10 -n 1 bash -c "$PATTERNS_DECL; check_url \"\$0\""
+}
+
+run_url_checks "$INPUT_FILE"
+
+if [ "$(cat "$TMP_FILE")" -eq 1 ]; then
+    echo "Errors found"
+    exit 1
+fi

scripts/check-modified.sh

@@ -0,0 +1,59 @@
+#!/bin/bash
+#
+# Checks for 404 links in a compiled list of modified books
+
+ERRORS=0
+
+FILES=$(git diff --name-only origin/main...HEAD --diff-filter=d -- "*.adoc")
+
+MODULES=$(echo "$FILES" | grep '^modules/.*\.adoc$')
+ASSEMBLIES=$(echo "$FILES" | grep '^assemblies/.*\.adoc$')
+BOOKS=$(echo "$FILES" | grep -E '^[^/]+\.adoc$')
+
+UPDATED_BOOKS=()
+
+if [ -n "$MODULES" ]; then
+  # Check for assemblies and books that include modified modules
+  while IFS= read -r module; do
+    mapfile -t updated_books < <(grep -rnwl . --include="*.adoc" --exclude-dir={_artifacts,modules,assemblies} -e "$(basename "$module")")
+    UPDATED_BOOKS+=( "${updated_books[@]}" )
+
+    mapfile -t updated_books < <(grep -rnwl assemblies --include="*.adoc" --exclude-dir={_artifacts,modules} -e "$(basename "$module")")
+    UPDATED_BOOKS+=( "${updated_books[@]}" )
+  done <<< "$MODULES"
+fi
+
+# Check for books that include modified assemblies
+if [ -n "$ASSEMBLIES" ]; then
+  while IFS= read -r assembly; do
+    mapfile -t results3 < <(grep -rnwl . --include="*.adoc" --exclude-dir={_artifacts,modules,assemblies} -e "$(basename "$assembly")")
+    UPDATED_BOOKS+=( "${results3[@]}" )
+  done <<< "$ASSEMBLIES"
+fi
+
+# Check for directly updated books
+if [ -n "$BOOKS" ]; then
+  while IFS= read -r book; do
+    UPDATED_BOOKS+=( "$book" )
+  done <<< "$BOOKS"
+fi
+
+if [ ${#UPDATED_BOOKS[@]} -eq 0 ]; then
+  echo "No modified books. Skipping link check."
+  exit 0
+fi
+
+# Check links in the compiled list of books
+
+for f in "${UPDATED_BOOKS[@]}"; do
+  echo "Checking: $f"
+  if ! ./scripts/check-links.sh "$f"; then
+    echo "❌ Link check failed for: $f"
+    ERRORS=1
+  fi
+done
+
+if [ "$ERRORS" -ne 0 ]; then
+  echo "One or more link checks failed."
+  exit 1
+fi

scripts/links.ignore

@@ -0,0 +1,15 @@
+# Add ignore link regexes one per line
+.*docs\.google\.com.*
+.*google\.com.*
+.*issues\.redhat\.com.*
+.*0\.0\.0\.0.*
+.*localhost.*
+.*registry\.redhat\.io.*
+.*example\.org.*
+.*github.com/example/myrepo\.git
+.*fonts\.googleapis\.com.*
+.*mixtral-my-project.apps\.my-cluster\.com.*
+.*openshiftapps\.com.*
+.*minio-cluster\.local.*
+.*codeflare-operator-webhook-service\.redhat-ods-applications\.svc
+.*example.com.*