blob: fbbc846b7be1a0ecc51bcc0806bb829de1d30452 [file] [log] [blame]
#!/usr/bin/env python3
# /// script
# requires-python = ">=3.10"
# dependencies = [
# "GitPython",
# "PyGithub",
# "rich-click",
# "packaging",
# "rich",
# ]
# ///
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
# This tool is based on the Spark merge_spark_pr script:
# https://github.com/apache/spark/blob/master/dev/merge_spark_pr.py
from __future__ import annotations
import re
import sys
from collections import Counter, defaultdict
from typing import TYPE_CHECKING
import git
import rich_click as click
from github import Auth, Github
from packaging import version
from rich.console import Console
from rich.progress import Progress
if TYPE_CHECKING:
from github.Issue import Issue
GIT_COMMIT_FIELDS = ["id", "author_name", "author_email", "date", "subject", "body"]
GIT_LOG_FORMAT = "%x1f".join(["%h", "%an", "%ae", "%ad", "%s", "%b"]) + "%x1e"
pr_title_re = re.compile(r".*\((#[0-9]{1,6})\)$")
STATUS_COLOR_MAP = {
"Closed": "yellow",
"Merged": "green",
"Open": "red",
}
DEFAULT_SECTION_NAME = "Uncategorized"
def get_commits_between(
repo,
previous_version,
target_version,
files=None,
):
log_args = [f"--format={GIT_LOG_FORMAT}", previous_version + ".." + target_version]
if files:
log_args.append("--")
log_args.append(" ".join(files))
log = repo.git.log(*log_args)
log = log.strip("\n\x1e").split("\x1e")
log = [row.strip().split("\x1f") for row in log]
return [dict(zip(GIT_COMMIT_FIELDS, row)) for row in log]
def style_issue_status(status):
if status in STATUS_COLOR_MAP:
return click.style(status[:6].ljust(6), STATUS_COLOR_MAP[status])
return status[:6].ljust(6)
def get_issue_type(issue):
label_prefix = "type:"
issue_type = DEFAULT_SECTION_NAME
if issue.labels:
for label in issue.labels:
if label.name.startswith(label_prefix):
return label.name.replace(label_prefix, "").strip()
if label.name == "changelog:skip":
return "(skip)"
return issue_type
def build_main_commits_cache(repo: git.Repo, issue_numbers: set[int]) -> dict[int, str]:
"""Build a cache of PR number -> main branch commit SHA with a single git log operation"""
cache = {}
if not issue_numbers:
return cache
try:
# Single git log to get all main branch commits
# Use a reasonable range to avoid scanning entire history
try:
# Try to get commits since last major version (should cover most PR ranges)
log_output = repo.git.log("origin/main", "--format=%H %s", "--since=1 year ago")
except Exception:
# Fallback to last 2000 commits if date fails
log_output = repo.git.log("origin/main", "--format=%H %s", "-2000")
# Use regex to find all PR numbers in commit messages at once
import re
pr_pattern = r"\(#(\d+)\)$" # PR number at end of commit message
for commit_line in log_output.splitlines():
if not commit_line:
continue
# Find PR number at end of commit message
match = re.search(pr_pattern, commit_line)
if match:
pr_number = int(match.group(1))
if pr_number in issue_numbers:
commit_sha = commit_line.split(" ")[0]
cache[pr_number] = commit_sha
except Exception:
# Fallback to empty cache if git operation fails
pass
return cache
def get_commit_in_main_associated_with_pr(repo: git.Repo, issue: Issue, main_commits_cache: dict[int, str]) -> str | None:
"""For a PR, find the associated merged commit & return its SHA"""
if issue.pull_request:
# Use cache (should be pre-populated)
return main_commits_cache.get(issue.number)
return None
def build_cherrypicked_cache(repo: git.Repo, issue_numbers: list[int], previous_version: str | None = None) -> dict[int, bool]:
"""Build a cache of which issues are cherry-picked by doing a single git log operation"""
cache = {num: False for num in issue_numbers}
if not issue_numbers:
return cache
# Get all commits in range and process them
log_args = ["--format=%H %s"]
if previous_version:
log_args.append(previous_version + "..")
try:
log_output = repo.git.log(*log_args)
# Use regex to find all PR numbers in the entire log output at once
import re
pr_pattern = r"\(#(\d+)\)"
for commit_line in log_output.splitlines():
# Find all PR numbers in this commit
matches = re.findall(pr_pattern, commit_line)
for match in matches:
issue_num = int(match)
if issue_num in cache:
cache[issue_num] = True
except Exception:
# Fallback to individual checks if batch fails
pass
return cache
def is_cherrypicked(repo: git.Repo, issue: Issue, previous_version: str | None = None) -> bool:
"""Check if a given issue is cherry-picked in the current branch or not"""
log_args = ["--format=%H %s", f"--grep=(#{issue.number})"]
if previous_version:
log_args.append(previous_version + "..")
log_output = repo.git.log(*log_args)
for commit_line in log_output.splitlines():
# Check if the PR number exists in the commit message (handles cherry-picked commits with multiple PR numbers)
if commit_line and f"(#{issue.number})" in commit_line:
return True
return False
def is_pr(issue: Issue) -> bool:
return "apache/airflow/pull/" in issue.html_url
def files_touched(repo, commit_id: str) -> list[str]:
real_commit = repo.commit(commit_id)
return real_commit.stats.files.keys()
def is_core_commit(files: list[str]) -> bool:
# We list out things that _aren't_ core files,
# and we want to know if the commit changes anything
# outside of these designated-non-core files
non_core_files = [
# Providers
"airflow/providers/",
"airflow/provider.yaml.schema.json",
"docs/apache-airflow-providers-",
# chart
"chart/",
"docs/helm-chart/",
"clients",
# non-released docs
"COMMITTERS.rst",
"contributing_docs/",
"INTHEWILD.md",
"INSTALL",
"README.md",
"images/",
"codecov.yml",
"kubernetes-tests/",
".github/",
".pre-commit-config.yaml",
"yamllint-config.yml",
".markdownlint.yml",
"tests/",
"dev/",
"Dockerfile",
"Dockerfile.ci",
".hadolint.yaml",
"scripts/",
"docs/build_docs.py",
"docs/start_doc_server.sh",
".devcontainer/",
"docker-context-files/",
# Misc
".dockerignore",
"docs/spelling_wordlist.txt",
"docs/integration-logos/*",
"docs/exts/",
"docs/docker-stack",
"docs/README.rst",
"docker-tests/",
"kubernetes-tests/",
"helm-tests/",
".asf.yaml",
".mailmap",
"breeze-legacy",
"breeze-complete",
".rat-excludes",
".gitattributes",
".gitpod.yml",
"generated/",
]
for file in files:
for ignore in non_core_files:
if file.startswith(ignore):
break
# Handle renaming. Form: {old_name => new_name}somename.py
if file.startswith("{"):
new_files = file[1:].split(" => ")
if any(n.strip().startswith(ignore) for n in new_files):
break
else:
return True
return False
def print_changelog(sections):
for section, lines in sections.items():
if section == "(skip)":
continue
print(section)
print('"' * len(section))
for line in lines:
print("-", line)
print()
@click.group()
def cli():
r"""
This tool should be used by Airflow Release Manager to verify what GitHub issues
were merged in the current working branch.
airflow-github compare <target_version> <github_token>
"""
@cli.command(short_help="Compare a GitHub target version against git merges")
@click.argument("target_version")
@click.argument("github-token", envvar="GITHUB_TOKEN")
@click.option(
"--previous-version",
"previous_version",
help="Specify the previous tag on the working branch to limit"
" searching for few commits to find the cherry-picked commits",
)
@click.option("--unmerged", "show_uncherrypicked_only", help="Show unmerged PRs only", is_flag=True)
@click.option("--show-commits", help="Show commit SHAs (default: on, off when --unmerged)", is_flag=True, default=None)
def compare(target_version, github_token, previous_version=None, show_uncherrypicked_only=False, show_commits=None):
# Set smart defaults
if show_commits is None:
show_commits = not show_uncherrypicked_only # Default off for --unmerged
repo = git.Repo(".", search_parent_directories=True)
github_handler = Github(auth=Auth.Token(github_token))
# Fetch PRs and Issues separately, with merged PRs identified upfront
merged_prs: list[Issue] = list(
github_handler.search_issues(
f'repo:apache/airflow milestone:"Airflow {target_version}" is:pull-request is:merged'
)
)
closed_prs: list[Issue] = list(
github_handler.search_issues(
f'repo:apache/airflow milestone:"Airflow {target_version}" is:pull-request is:closed -is:merged'
)
)
open_prs: list[Issue] = list(
github_handler.search_issues(
f'repo:apache/airflow milestone:"Airflow {target_version}" is:pull-request is:open'
)
)
# Skip fetching issues if we only care about unmerged PRs
if show_uncherrypicked_only:
issues = []
else:
issues: list[Issue] = list(
github_handler.search_issues(
f'repo:apache/airflow milestone:"Airflow {target_version}" is:issue'
)
)
# Create a merge status lookup
pr_merge_status_cache = {}
for pr in merged_prs:
pr_merge_status_cache[pr.number] = True
for pr in closed_prs:
pr_merge_status_cache[pr.number] = False
# Open PRs are neither merged nor closed, so we don't need to cache them
milestone_issues = merged_prs + closed_prs + open_prs + issues
num_cherrypicked = 0
num_uncherrypicked = Counter()
# :<18 says left align, pad to 18, :>6 says right align, pad to 6
# :<50.50 truncates after 50 chars
# !s forces as string
if show_commits:
formatstr = (
"{number:>6} | {typ!s:<5} | {changelog!s:<13} | {status!s} "
"| {title:<83.83} | {merged:<6} | {commit:>7.7} | {url}"
)
header_fields = {
"number": "NUMBER",
"typ": "TYPE",
"changelog": "CHANGELOG",
"status": "STATUS".ljust(6),
"title": "TITLE",
"merged": "CHERRY",
"commit": "COMMIT",
"url": "URL",
}
else:
formatstr = (
"{number:>6} | {typ!s:<5} | {changelog!s:<13} | {status!s} "
"| {title:<95.95} | {merged:<6} | {url}"
)
header_fields = {
"number": "NUMBER",
"typ": "TYPE",
"changelog": "CHANGELOG",
"status": "STATUS".ljust(6),
"title": "TITLE",
"merged": "CHERRY",
"commit": "", # Not used
"url": "URL",
}
print(formatstr.format(**header_fields))
milestone_issues = sorted(
milestone_issues, key=lambda x: x.closed_at if x.closed_at else x.created_at, reverse=True
)
# Build caches for performance optimization
issue_numbers = [issue.number for issue in milestone_issues if is_pr(issue)]
# Convert to set for O(1) lookups in cache building
issue_numbers_set = set(issue_numbers)
# Build all caches upfront with batch operations
if show_commits:
main_commits_cache = build_main_commits_cache(repo, issue_numbers_set)
else:
main_commits_cache = {}
cherrypicked_cache = build_cherrypicked_cache(repo, issue_numbers, previous_version)
for issue in milestone_issues:
issue_is_pr = is_pr(issue)
# Determine status - differentiate between Closed and Merged for PRs using cache
if issue_is_pr and issue.state == "closed":
is_merged = pr_merge_status_cache.get(issue.number, False)
status = "Merged" if is_merged else "Closed"
else:
status = issue.state.capitalize()
# Checks if commit was cherrypicked into branch using cache
if issue_is_pr:
is_cherry_picked = cherrypicked_cache.get(issue.number, False)
else:
is_cherry_picked = is_cherrypicked(repo, issue, previous_version)
if is_cherry_picked:
num_cherrypicked += 1
if show_uncherrypicked_only:
continue
cherrypicked = click.style("Yes".ljust(6), "green")
elif not issue_is_pr and show_uncherrypicked_only:
# Don't show issues when looking for unmerged PRs
continue
elif issue_is_pr:
num_uncherrypicked[status] += 1
cherrypicked = click.style("No".ljust(6), "red")
else:
cherrypicked = ""
fields = dict(
number=issue.number,
typ="PR" if issue_is_pr else "Issue",
changelog=get_issue_type(issue) if issue_is_pr else "",
status=style_issue_status(status),
title=issue.title,
url=issue.html_url,
)
# Only get commit info if we're showing commits
if show_commits:
commit_in_main = get_commit_in_main_associated_with_pr(repo, issue, main_commits_cache)
fields["commit"] = commit_in_main if commit_in_main else ""
print(formatstr.format(**fields, merged=cherrypicked))
print(
f"Commits on branch: {num_cherrypicked:d}, {sum(num_uncherrypicked.values()):d} "
f"({dict(num_uncherrypicked)}) yet to be cherry-picked"
)
@cli.command(short_help="Build a CHANGELOG grouped by GitHub Issue type")
@click.argument("previous_version")
@click.argument("target_version")
@click.argument("github-token", envvar="GITHUB_TOKEN")
@click.option("--disable-progress-bar", is_flag=True, help="Disable the progress bar")
def changelog(previous_version, target_version, github_token, disable_progress_bar):
repo = git.Repo(".", search_parent_directories=True)
# Get a list of issues/PRs that have been committed on the current branch.
log = get_commits_between(repo, previous_version, target_version)
if disable_progress_bar:
print(f"Processing {len(log)} commits")
gh = Github(auth=Auth.Token(github_token))
gh_repo = gh.get_repo("apache/airflow")
sections = defaultdict(list)
with open("RELEASE_NOTES.rst") as file:
existing_pr_nos = re.findall(r"\(#(\d+)\)", file.read())
console = Console(width=180)
with Progress(console=console, disable=disable_progress_bar) as progress:
task = progress.add_task("Processing commits from changelog", total=len(log))
for commit in progress.track(log, description="Processing commits from changelog"):
tickets = pr_title_re.findall(commit["subject"])
if tickets:
issue = gh_repo.get_issue(number=int(tickets[0][1:]))
issue_type = get_issue_type(issue)
files = files_touched(repo, commit["id"])
if str(issue.number) in existing_pr_nos:
print(f"Skipping {issue.number} as it's already in RELEASE_NOTES.rst")
continue
if is_core_commit(files):
sections[issue_type].append(commit["subject"])
progress.update(task, advance=1)
else:
sections[DEFAULT_SECTION_NAME].append(commit["subject"])
progress.update(task, advance=1)
console.print("\n")
print_changelog(sections)
@cli.command(short_help="Find merged PRs that still need to be categorized for the changelog")
@click.argument("previous_version")
@click.argument("target_version")
@click.option("--show-skipped", is_flag=True)
@click.option("--show-files", is_flag=True)
@click.argument("github-token", envvar="GITHUB_TOKEN")
def needs_categorization(previous_version, target_version, show_skipped, show_files, github_token):
repo = git.Repo(".", search_parent_directories=True)
log = get_commits_between(repo, previous_version, target_version)
gh = Github(auth=Auth.Token(github_token))
gh_repo = gh.get_repo("apache/airflow")
for commit in log:
tickets = pr_title_re.findall(commit["subject"])
if tickets:
issue = gh_repo.get_issue(number=int(tickets[0][1:]))
issue_type = get_issue_type(issue)
if issue_type == DEFAULT_SECTION_NAME:
files = files_touched(repo, commit["id"])
if is_core_commit(files):
print(f"{commit['subject']}: {issue.html_url}")
if show_files:
for f in files:
print(f"\t{f}")
elif show_skipped:
print(f"**** {commit['subject']}: Skipping - not a core commit")
if show_files:
for f in files:
print(f"\t{f}")
else:
print(f"Commit '{commit['id']}' is missing PR number: {commit['subject']}")
@cli.command(
name="api-clients-policy",
help="Compare two airflow core release tags and determine if API clients need to be released.",
)
@click.argument("previous_version")
@click.argument("target_version")
def api_clients_policy(previous_version, target_version):
p_version = version.parse(previous_version)
t_version = version.parse(target_version)
if p_version.major != t_version.major:
print("This is a major release, API clients should also be released.")
return
if p_version.minor != t_version.minor:
print("This is a minor release, API clients should also be released.")
return
if p_version == t_version:
print("Both versions are identical")
return
repo = git.Repo(".", search_parent_directories=True)
log = get_commits_between(
repo,
previous_version,
target_version,
files=[f"{repo.working_dir}/airflow-core/src/airflow/api_fastapi/core_api/openapi/v2-rest-api-generated.yaml"],
)
clients_need_release = False
for commit in log:
if "update airflow version to" not in commit["subject"].lower():
clients_need_release = True
print(f"Commit '{commit['id']}' updated the OpenAPI spec, PR number: {commit['subject']}")
if clients_need_release:
print(f"API clients need to be released because the API spec has changed since '{previous_version}'")
else:
print(
"API clients don't need to be released because the API spec hasn't changed between those two "
"patch versions"
)
if __name__ == "__main__":
import doctest
(failure_count, test_count) = doctest.testmod()
if failure_count:
sys.exit(-1)
try:
cli()
except Exception:
raise