ci: respect redirected pages in removed URL check

.github/workflows/check-removed-urls.yml

@@ -36,20 +36,17 @@ jobs:
       - name: Generate current URLs list
         run: |
           for dir in compare base; do
-            pushd ${dir}/docs
-            find ./_build/ -name '*.html' \
-            | sed 's|/_build||;s|/index.html$|/|;s|.html$||' \
-            | sort > urls.txt
-            popd
+            build_dir="${dir}/docs/_build"
+            urls_file="${dir}/docs/urls.txt"
+
+            if [ ! -d "${build_dir}" ]; then
+              echo "Expected docs build directory not found: ${build_dir}"
+              exit 1
+            fi
+
+            find "${build_dir}" -name '*.html' \
+              | sed "s|^${build_dir}||;s|^/html||;s|/index.html$|/|;s|.html$||" \
+              | sort > "${urls_file}"
           done
       - name: Compare URLs
-        run: |
-          BASE_URLS_PATH="base/docs/urls.txt"
-          COMPARE_URLS_PATH="compare/docs/urls.txt"
-          removed=$(comm -23 ${BASE_URLS_PATH} ${COMPARE_URLS_PATH} )
-          if [ -n "$removed" ]; then
-            echo "The following URLs were removed:"
-            echo "$removed"
-            echo "Please ensure removed pages are redirected"
-            exit 1
-          fi
+        run: python3 compare/docs/_dev/check_removed_urls.py

CHANGELOG.md

@@ -2,6 +2,9 @@
 
 ## Upcoming
 
+* Prevent Vale from processing Markdown files in the build directory
+* Update link to documenation in README
+* Make removed URL check redirect-aware and add support for reusable workflow contexts
 * Exclude utility directories from builds and checks
 * Update link to documentation in README
 
@@ -11,6 +14,7 @@
 * `docs/Makefile` [#605](https://github.com/canonical/sphinx-stack/pull/605), [#610](https://github.com/canonical/sphinx-stack/pull/610)
 * `README.md` [#603](https://github.com/canonical/sphinx-stack/pull/603)
 * `.github/workflows/cla-check.yml` [#606](https://github.com/canonical/sphinx-stack/pull/606)
+* `.github/workflows/check-removed-urls.yml` [#PR_NUMBER](https://github.com/canonical/sphinx-stack/pull/#PR_NUMBER)
 
 ## 2.0
 

docs/_dev/check_removed_urls.py

@@ -0,0 +1,98 @@
+#! /usr/bin/env python
+
+"""Check for removed URLs and verify if redirects exist."""
+
+import csv
+import io
+import sys
+from pathlib import Path
+
+BASE_URLS = Path("base/docs/urls.txt")
+COMPARE_URLS = Path("compare/docs/urls.txt")
+REDIRECTS = Path("compare/docs/redirects.txt")
+
+
+def read_urls(path):
+    return {
+        line.strip()
+        for line in path.read_text(encoding="utf-8").splitlines()
+        if line.strip()
+    }
+
+
+def read_redirect_sources(path):
+    sources = set()
+
+    if not path.exists():
+        return sources
+
+    for raw_line in path.read_text(encoding="utf-8").splitlines():
+        line = raw_line.strip()
+        if not line or line.startswith("#"):
+            continue
+
+        fields = next(
+            csv.reader(
+                io.StringIO(line),
+                delimiter=" ",
+                quotechar='"',
+                skipinitialspace=True,
+            ),
+            [],
+        )
+        if fields:
+            sources.add(fields[0])
+
+    return sources
+
+
+def source_candidates_for_url(url):
+    clean_path = url.strip()
+    clean_path = clean_path.removeprefix("./")
+    clean_path = clean_path.removeprefix("/")
+    clean_path = clean_path.removesuffix(".html")
+    clean_path = clean_path.rstrip("/")
+
+    if not clean_path:
+        return {"index.md"}
+
+    # A removed dirhtml URL can map back to either a page file or an
+    # index file. Directory-level redirects are stored with a trailing
+    # slash, so include that form too.
+    return {
+        f"{clean_path}.md",
+        f"{clean_path}/index.md",
+        f"{clean_path}/",
+    }
+
+
+def main():
+    if not BASE_URLS.exists():
+        print(f"Error: Base URLs file not found at {BASE_URLS}")
+        sys.exit(1)
+    if not COMPARE_URLS.exists():
+        print(f"Error: Compare URLs file not found at {COMPARE_URLS}")
+        sys.exit(1)
+
+    removed_urls = sorted(read_urls(BASE_URLS) - read_urls(COMPARE_URLS))
+    redirect_sources = read_redirect_sources(REDIRECTS)
+
+    missing_redirects = [
+        url
+        for url in removed_urls
+        if source_candidates_for_url(url).isdisjoint(redirect_sources)
+    ]
+
+    if missing_redirects:
+        print("The following URLs were removed without redirects:")
+        print("\n".join(missing_redirects))
+        print("Please ensure removed pages are redirected")
+        sys.exit(1)
+
+    if removed_urls:
+        print("Removed URLs have redirects:")
+        print("\n".join(removed_urls))
+
+
+if __name__ == "__main__":
+    main()