Merge pull request #64807 from ClickHouse/adjust-changelogs

Migrate changelog.py to a descendant of fuzzywuzzy

docker/test/style/Dockerfile

@@ -15,7 +15,6 @@ RUN apt-get update && env DEBIAN_FRONTEND=noninteractive apt-get install --yes \
         file \
         libxml2-utils \
         moreutils \
-        python3-fuzzywuzzy \
         python3-pip \
         yamllint \
         locales \
@@ -23,8 +22,18 @@ RUN apt-get update && env DEBIAN_FRONTEND=noninteractive apt-get install --yes \
     && rm -rf /var/lib/apt/lists/* /var/cache/debconf /tmp/*
 
 # python-magic is the same version as in Ubuntu 22.04
-RUN pip3 install black==23.12.0 boto3 codespell==2.2.1 mypy==1.8.0 PyGithub unidiff pylint==3.1.0 \
-        python-magic==0.4.24 requests types-requests \
+RUN pip3 install \
+        PyGithub \
+        black==23.12.0 \
+        boto3 \
+        codespell==2.2.1 \
+        mypy==1.8.0 \
+        pylint==3.1.0 \
+        python-magic==0.4.24 \
+        requests \
+        thefuzz \
+        types-requests \
+        unidiff \
     && rm -rf /root/.cache/pip
 
 RUN echo "en_US.UTF-8 UTF-8" > /etc/locale.gen && locale-gen en_US.UTF-8

utils/changelog/README.md

@@ -6,7 +6,7 @@ Generate github token:
 Dependencies:
 ```
 sudo apt-get update
-sudo apt-get install git python3 python3-fuzzywuzzy python3-github
+sudo apt-get install git python3 python3-thefuzz python3-github
 python3 changelog.py -h
 ```
 
@@ -15,10 +15,7 @@ Usage example:
 Note: The working directory is ClickHouse/utils/changelog
 
 ```bash
-export GITHUB_TOKEN="<your token>"
-
-git fetch --tags # changelog.py depends on having the tags available, this will fetch them.  
-                 # If you are working from a branch in your personal fork, then you may need `git fetch --all`
+GITHUB_TOKEN="<your token>"
 
 python3 changelog.py --output=changelog-v22.4.1.2305-prestable.md --gh-user-or-token="$GITHUB_TOKEN" v21.6.2.7-prestable
 python3 changelog.py --output=changelog-v22.4.1.2305-prestable.md --gh-user-or-token="$USER" --gh-password="$PASSWORD" v21.6.2.7-prestable

utils/changelog/changelog.py

@@ -10,9 +10,9 @@ from datetime import date, timedelta
 from subprocess import DEVNULL, CalledProcessError
 from typing import Dict, List, Optional, TextIO
 
-from fuzzywuzzy.fuzz import ratio  # type: ignore
 from github.GithubException import RateLimitExceededException, UnknownObjectException
 from github.NamedUser import NamedUser
+from thefuzz.fuzz import ratio  # type: ignore
 
 from git_helper import git_runner as runner
 from git_helper import is_shallow
@@ -100,17 +100,14 @@ def get_descriptions(prs: PullRequests) -> Dict[str, List[Description]]:
         # obj._rawData doesn't spend additional API requests
         # We'll save some requests
         # pylint: disable=protected-access
-        repo_name = pr._rawData["base"]["repo"]["full_name"]  # type: ignore
+        repo_name = pr._rawData["base"]["repo"]["full_name"]
         # pylint: enable=protected-access
         if repo_name not in repos:
             repos[repo_name] = pr.base.repo
         in_changelog = False
         merge_commit = pr.merge_commit_sha
-        try:
-            runner.run(f"git rev-parse '{merge_commit}'")
-        except CalledProcessError:
-            # It's possible that commit not in the repo, just continue
-            logging.info("PR %s does not belong to the repo", pr.number)
+        if merge_commit is None:
+            logging.warning("PR %s does not have merge-commit, skipping", pr.number)
             continue
 
         in_changelog = merge_commit in SHA_IN_CHANGELOG

utils/changelog/requirements.txt

@@ -1,3 +1,2 @@
-fuzzywuzzy
+thefuzz
 PyGitHub
-python-Levenshtein