Add workflow to check redirects for renamed files

.github/workflows/redirect_checks.yaml

@@ -0,0 +1,57 @@
+name: "Redirects in renamed files"
+"on":
+  pull_request:
+    branches:
+      - main
+
+jobs:
+  prose:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Get changed files in the docs folder
+        id: changed-files
+        uses: tj-actions/changed-files@v34
+        with:
+          include_all_old_new_renamed_files: true
+          files: |
+            docs/**
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: 3.8
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install -r requirements-dev.txt
+
+      - name: Check renamed files
+        id: missing_redirection
+        env:
+          ALL_OLD_AND_NEW_RENAMED_FILES: ${{ steps.changed-files.outputs.all_old_new_renamed_files }}
+        run: |
+          for file in ${{ steps.changed-files.outputs.renamed_files }};
+          do
+            if [ "${file: -4}" == ".rst" ]
+              file="$file" | sed -e 's/\.[^.]*$//'
+              then
+                var="$var $file"
+            fi
+          done
+          echo "$var"
+          if ! [ -z "$var" ];then
+            python "scripts/check_renamed_files.py" --renamed_files ${{ steps.changed-files.outputs.renamed_files }}
+          fi
+
+      - name: Check env
+        env:
+          MISSING_REDIRECTION: ${{ env.MISSING_REDIRECTION }}
+        if: "${{ env.MISSING_REDIRECTION != '' }}"
+        run: |
+          echo "::error::Files were renamed with missing _redirects. Add redirect to renamed files to fix it. Renamed: ${{env.MISSING_REDIRECTION}}"
+          echo "::error::For full report, check logs on the step above."
+          exit 1

requirements-dev.txt

@@ -1 +1,2 @@
 ultimate-sitemap-parser==0.5
+coloredlogs==15.0.1

scripts/check_renamed_files.py

@@ -0,0 +1,123 @@
+"""
+This script checks if a renamed file is added for proper redirection on redirects_ file.
+"""
+import os, sys, argparse
+import coloredlogs, logging
+from typing import Dict, List
+
+myPath = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, myPath + "/../../")
+
+
+logger = logging.getLogger("check_renamed_files")
+logging.basicConfig(format="%(asctime)s - [%(levelname)s] - %(message)s")
+
+
+FIELD_STYLES = dict(asctime=dict(color="yellow"), levelname=dict(color="magenta"))
+LEVEL_STYLES = dict(
+    debug=dict(color="green"),
+    info=dict(color="cyan"),
+    verbose=dict(color="blue"),
+    warning=dict(color="yellow"),
+    error=dict(color="red"),
+    critical=dict(color="red"),
+)
+
+coloredlogs.install(
+    level="DEBUG",
+    logger=logger,
+    isatty=True,
+    level_styles=LEVEL_STYLES,
+    field_styles=FIELD_STYLES,
+)
+
+
+def find_redirected() -> Dict:
+    """Creates a dict with the redirected files
+    on the _redirects file.
+
+    { previous : redirected}
+
+    :returns: set with redirected files
+    :rtype: set
+    """
+    REDIRECTED_FILE = "_redirects"
+    all_redirected_links = {}
+    with open(REDIRECTED_FILE) as f:
+        lines = f.readlines()
+    for line in lines:
+        if line.startswith("/docs/") and not line.startswith("/docs/:"):
+            previous, redirected = [x.lstrip("/") for x in line.split()]
+            # avoid repetitions of `html`
+            if not previous.endswith(".html"):
+                all_redirected_links[previous] = redirected
+
+    return all_redirected_links
+
+
+def check_missing_redirects(renamed_files: List[str]):
+    """
+    Creates a dict with all missing redirects.
+
+    { previous_link: current_link }
+
+    :returns: missing redirects
+    :rtype: dict
+    """
+
+    INPUT = os.environ["ALL_OLD_AND_NEW_RENAMED_FILES"]
+    all_new_and_renamed_files = dict([x.split(",")[::-1] for x in INPUT.split(" ")])
+
+    missing_redirects = {}
+    all_redirected_links = find_redirected()
+    for renamed in renamed_files:
+        try:
+            previous_link = all_new_and_renamed_files[renamed].rstrip(".rst")
+            logger.debug(previous_link)
+        except KeyError:
+            logger.error("Missing renamed files on all new and renamed files.")
+            exit(0)
+        current_link = renamed.rstrip(".rst")
+        if all_redirected_links.get(previous_link) != current_link:
+            missing_redirects[previous_link] = current_link
+    return missing_redirects
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description="Script to find missing redirects.")
+    parser.add_argument(
+        "--renamed_files",
+        help="delimited list input",
+        type=str,
+        nargs="*",
+        default=" ",
+    )
+    parser.add_argument(
+        "--path_file", help="path to a file", type=str, default=os.getenv("GITHUB_ENV")
+    )
+
+    renamed_files = parser.parse_args().renamed_files
+    env_file = parser.parse_args().path_file
+
+    if not renamed_files:
+        logger.info("No files are renamed.")
+        sys.exit()
+
+    missing_redirects = check_missing_redirects(renamed_files)
+    if not missing_redirects:
+        logger.info("Files renamed and redirects are added to _redirects.")
+        sys.exit()
+
+    res = ""
+    h = ["Previous Name", " Current Name"]
+    logger.error("{:<45s} {:<45s}".format(*h))
+    for k, v in missing_redirects.items():
+        logger.error("/{:<45s} /{:<45s}".format(k, v))
+
+    logger.info("🚨 Seems like you forgot to add redirects for the renamed files. 🚨")
+    logger.info(
+        "ℹ️ More info: https://docs.aiven.io/docs/community/documentation/tips-tricks/renaming-files.html"
+    )
+
+    with open(env_file, "a") as myfile:
+        myfile.write(f"MISSING_REDIRECTION={missing_redirects}")

scripts/tests/test_renamed_files.py

@@ -0,0 +1,58 @@
+import sys, os
+import mock, pytest
+
+myPath = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, myPath + "/../../")
+from scripts.check_renamed_files import check_missing_redirects
+
+
+@mock.patch.dict(
+    os.environ,
+    {
+        "ALL_OLD_AND_NEW_RENAMED_FILES": "docs/tools.rst,docs/renamed_tools_file.rst docs/community.rst,docs/renamed_community_file.rst docs/integrations.rst,docs/renamed_integration_file.rst docs/platform.rst,docs/platform2.rst docs/products/opensearch/howto/connect-with-python.rst,docs/products/opensearch/howto/renamed_opensearch_file.rst"
+    },
+)
+def test_check_missing_redirects_with_missing_links():
+    renamed_files = [
+        "docs/renamed_community_file.rst",
+        "docs/renamed_integration_file.rst",
+        "docs/renamed_tools_file.rst",
+        "docs/products/opensearch/howto/renamed_opensearch_file.rst",
+    ]
+    missing_redirects = check_missing_redirects(renamed_files)
+    assert missing_redirects == {
+        "docs/community": "docs/renamed_community_file",
+        "docs/integration": "docs/renamed_integration_file",
+        "docs/tool": "docs/renamed_tools_file",
+        "docs/products/opensearch/howto/connect-with-python": "docs/products/opensearch/howto/renamed_opensearch_file",
+    }
+
+
+@mock.patch.dict(
+    os.environ,
+    {
+        "ALL_OLD_AND_NEW_RENAMED_FILES": "docs/tools.rst,docs/renamed_tools_file.rst docs/community.rst,docs/renamed_community_file.rst docs/integrations.rst,docs/renamed_integration_file.rst docs/platform.rst,docs/platform2.rst docs/products/opensearch/howto/connect-with-python.rst,docs/products/opensearch/howto/renamed_opensearch_file.rst"
+    },
+)
+def test_check_missing_redirects_with_no_renamed_files():
+    """
+    Test when there were previous changes in all old and new renamed file,
+    without renamed files in current commit. No missind redirects should be
+    reported
+    """
+    renamed_files = []
+    missing_redirects = check_missing_redirects(renamed_files)
+    assert missing_redirects == {}
+
+
+@mock.patch.dict(
+    os.environ,
+    {"ALL_OLD_AND_NEW_RENAMED_FILES": ""},
+)
+def test_check_missing_redirects_no_files_changed_and_no_renamed_files():
+    """
+    Function should raise an error if no files changed and this function is called
+    """
+    with pytest.raises(ValueError):
+        renamed_files = []
+        check_missing_redirects(renamed_files)