Changelog automation [WIP]

tools/update_changelog.py

@@ -0,0 +1,248 @@
+#!/usr/bin/env python3
+
+"""
+Git Extract Changelog Entries
+
+This script iterates through Git commits and extracts a changelog line for all
+merged pull request in the given range to stdout
+
+Usage:
+    python script.py --start_commit <start_commit> \
+    [--end_commit <end_commit>] [--repo_path /path/to/your/repo]
+
+Dependencies:
+    - GitPython library
+"""
+
+import git
+import requests
+import re
+import os
+import argparse
+import collections
+
+r = requests.Session()
+
+
+def get_category_key(category):
+    return category if category != "Uncategorized" else None
+
+
+def get_milestone_key(milestone):
+    return milestone.get("title") if milestone else None
+
+
+def get_milestone_title(milestone):
+    return (
+        f"[{milestone.get('title')}]({milestone.get('url')})"
+        if milestone
+        else "Unreleased"
+    )
+
+
+class Milestone(dict):
+    def __init__(self):
+        super().__init__()
+        self.start = None
+
+    def offset(self, threshold, lines):
+        for k, v in self.items():
+            if v >= threshold:
+                self[k] += lines
+        if self.start >= threshold:
+            self.start += lines
+
+
+class Changelog:
+    def __init__(self):
+        self.keys = collections.defaultdict(Milestone)
+        self.start = None
+        self.lines = []
+        self.pr_numbers = set()
+
+    def offset(self, threshold, lines=1):
+        for v in self.keys.values():
+            v.offset(threshold, lines)
+        if self.start >= threshold:
+            self.start += lines
+
+    def read(self, f):
+        current_milestone = None
+        current_category = None
+        pr_pattern = re.compile(r"\[#(\d+)\]")
+        version = re.compile(r"\[(\d\.\d\.\d)\]")
+        self.pr_numbers.clear()
+
+        for nth, line in enumerate(f.readlines()):
+            self.lines.append(line)
+            match = pr_pattern.search(line)
+            if line.startswith("## "):
+                if not current_milestone:
+                    self.start = nth - 1
+                if current_category:
+                    self.keys[current_milestone][
+                        get_category_key(current_category)
+                    ] = (nth - 1)
+                    current_category = None
+                match = version.search(line)
+                if match:
+                    current_milestone = match.group(1)
+                elif line.strip().endswith("Unreleased"):
+                    current_milestone = None
+                else:
+                    current_milestone = line[2:].strip()
+                self.keys[current_milestone].start = nth + 1
+            elif line.startswith("### "):
+                if current_category:
+                    self.keys[current_milestone][
+                        get_category_key(current_category)
+                    ] = (nth - 1)
+                current_category = line[3:].strip()
+            elif match:
+                self.pr_numbers.add(match.group(1))
+
+    def add(self, milestone, category, item):
+        key = get_milestone_key(milestone)
+        if key not in self.keys:
+            self.lines.insert(
+                self.start, f"\n## {get_milestone_title(milestone)}\n"
+            )
+            self.offset(self.start)
+            self.keys[key].start = self.start
+        if category not in self.keys[key]:
+            self.lines.insert(
+                self.keys[key].start,
+                (
+                    f"\n### {category}\n\n"
+                    if category
+                    else "\n### Uncategorized\n\n"
+                ),
+            )
+            self.offset(self.keys[key].start)
+            self.keys[key][category] = self.keys[key].start
+            if key is None:
+                self.keys[key].start += 1
+        self.lines.insert(self.keys[key][category], item)
+        self.offset(self.keys[key][category])
+
+    def write(self, f):
+        f.writelines(self.lines)
+
+
+def fetch_info(owner, repo, pull_number):
+    query = """query GetProjectItems($owner: String!, $repo: String!, $pull_number: Int!) {
+        repository(owner: $owner, name: $repo) {
+            pullRequest(number: $pull_number) {
+                projectItems(first: 10) {
+                    nodes {
+                        fieldValueByName(name: "Category") {
+                            ... on ProjectV2ItemFieldSingleSelectValue {
+                                name
+                                field {
+                                ... on ProjectV2SingleSelectField {
+                                    name
+                                }
+                            }
+                        }
+                    }
+                }
+            }
+            milestone {
+                title
+                url
+            }
+            labels(first: 20) {
+                nodes {
+                    name
+                }
+            }
+        }
+    }
+    }"""
+    response = r.post(
+        "https://api.github.com/graphql",
+        json=dict(
+            query=query,
+            variables=dict(owner=owner, repo=repo, pull_number=pull_number),
+        ),
+        headers=dict(Authorization=f"Bearer {os.getenv('GITHUB_TOKEN')}"),
+    )
+    data = (
+        response.json()
+        .get("data", {})
+        .get("repository", {})
+        .get("pullRequest", {})
+    )
+    category = None
+    labels = []
+    nodes = data.get("projectItems", {}).get("nodes", [])
+    if nodes:
+        category = nodes[0].get("fieldValueByName", {}).get("name")
+    milestone = data.get("milestone")
+    nodes = data.get("labels")
+    if nodes:
+        labels = [
+            node.get("name")
+            for node in nodes.get("nodes", [])
+            if node.get("name")
+        ]
+    return category, milestone, labels
+
+
+changelog = Changelog()
+
+
+def extract_pull_requests(repo_path, end_commit, file):
+    repo = git.Repo(repo_path)
+    start_commit = next(repo.iter_commits(paths=file, max_count=1))
+
+    pr_pattern = re.compile(r"\[#(\d+)\]")
+    with open(file) as f:
+        changelog.read(f)
+
+    pr_pattern = re.compile(r"Merge pull request #(\d+)")
+
+    for commit in repo.iter_commits(rev=f"{start_commit}..{end_commit}"):
+        match = pr_pattern.search(commit.summary)
+        if match:
+            pr_number = match.group(1)
+            commit_message_lines = commit.message.split("\n")
+
+            if pr_number in changelog.pr_numbers:
+                continue
+
+            category, milestone, labels = fetch_info(
+                owner="mixxxdj", repo="mixxx", pull_number=int(pr_number)
+            )
+            if "sync-branches" in labels or "changelog" in labels:
+                continue
+            changelog.add(
+                milestone,
+                category,
+                f"* {commit_message_lines[2]} [#{pr_number}](https://github.com/mixxxdj/mixxx/pull/{pr_number})\n",
+            )
+
+    with open(file, "w") as f:
+        changelog.write(f)
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(
+        description="Git Extract Changelog Entries"
+    )
+    parser.add_argument(
+        "--end_commit", default="HEAD", help="Ending commit SHA (default HEAD)"
+    )
+    parser.add_argument(
+        "--repo_path",
+        default=".",
+        help="Path to the Git repository (default: current directory)",
+    )
+    parser.add_argument(
+        "--file",
+        default=".",
+        help="Path to the changelog file (default: CHANGELOG.md)",
+    )
+
+    args = parser.parse_args()
+    extract_pull_requests(args.repo_path, args.end_commit, args.file)