List PRs in changelog generator

.github/generate_changelog.py

@@ -8,7 +8,8 @@ import re
 import sys
 import json
 import argparse
-from typing import List, Tuple, Optional
+import os
+from typing import List, Tuple, Optional, Dict, Any
 
 
 def run_git_command(args: List[str]) -> str:
@@ -55,46 +56,132 @@ def get_submodule_commit(submodule_path: str, tag: str) -> Optional[str]:
         return None
 
 
-def get_commits_between_tags(tag1: str, tag2: str) -> List[Tuple[str, str]]:
+def get_repo_info() -> Tuple[str, str]:
-    """Get commits between two tags. Returns list of (message, author) tuples."""
+    """Get repository owner and name from git remote."""
+    try:
+        remote_url = run_git_command(["config", "--get", "remote.origin.url"])
+
+        # Handle both HTTPS and SSH URLs
+        # SSH: git@github.com:owner/repo.git
+        # HTTPS: https://github.com/owner/repo.git
+        match = re.search(r'github\.com[:/](.+?)/(.+?)(?:\.git)?$', remote_url)
+        if match:
+            owner = match.group(1)
+            repo = match.group(2)
+            return owner, repo
+        else:
+            print("Error: Could not parse GitHub repository from remote URL", file=sys.stderr)
+            sys.exit(1)
+    except:
+        print("Error: Could not get git remote URL", file=sys.stderr)
+        sys.exit(1)
+
+
+def get_commits_between_tags(tag1: str, tag2: str) -> List[str]:
+    """Get commit SHAs between two tags."""
     log_output = run_git_command([
         "log",
         f"{tag2}..{tag1}",
-        "--format=%s|%an|%h"
+        "--format=%H"
     ])
 
-    commits = []
+    commits = [sha.strip() for sha in log_output.split("\n") if sha.strip()]
-    for line in log_output.split("\n"):
-        if "|" in line:
-            message, author, sha = line.split("|", 2)
-            commits.append((message.strip(), author.strip(), sha.strip()))
-    
     return commits
 
 
-def filter_commits(commits: List[Tuple[str, str]], ignore_patterns: List[str]) -> List[Tuple[str, str]]:
+def get_pr_for_commit(commit_sha: str, owner: str, repo: str, token: str) -> Optional[Dict[str, Any]]:
-    """Filter out commits matching any of the ignore patterns."""
+    """Get PR information for a commit using GitHub API."""
+    try:
+        import requests
+    except ImportError:
+        print("Error: requests library is required. Install it with: pip install requests", file=sys.stderr)
+        sys.exit(1)
+
+    headers = {
+        "Accept": "application/vnd.github+json",
+        "X-GitHub-Api-Version": "2022-11-28"
+    }
+
+    if token:
+        headers["Authorization"] = f"Bearer {token}"
+
+    url = f"https://api.github.com/repos/{owner}/{repo}/commits/{commit_sha}/pulls"
+
+    try:
+        response = requests.get(url, headers=headers)
+        response.raise_for_status()
+        prs = response.json()
+
+        if prs and len(prs) > 0:
+            # Return the first PR (most relevant one)
+            pr = prs[0]
+            return {
+                "number": pr["number"],
+                "title": pr["title"],
+                "author": pr["user"]["login"],
+                "url": pr["html_url"]
+            }
+    except requests.exceptions.HTTPError as e:
+        if e.response.status_code == 404:
+            # Commit might not be associated with a PR
+            return None
+        elif e.response.status_code == 403:
+            print("Warning: GitHub API rate limit exceeded. Consider providing a token.", file=sys.stderr)
+            return None
+        else:
+            print(f"Warning: Failed to fetch PR for commit {commit_sha[:7]}: {e}", file=sys.stderr)
+            return None
+    except Exception as e:
+        print(f"Warning: Error fetching PR for commit {commit_sha[:7]}: {e}", file=sys.stderr)
+        return None
+
+    return None
+
+
+def get_prs_between_tags(tag1: str, tag2: str, owner: str, repo: str, token: str) -> List[Dict[str, Any]]:
+    """Get PRs between two tags using GitHub API."""
+    commits = get_commits_between_tags(tag1, tag2)
+    print(f"Found {len(commits)} commits, fetching PR information...")
+
+    prs = []
+    seen_pr_numbers = set()
+
+    for i, commit_sha in enumerate(commits, 1):
+        if i % 10 == 0:
+            print(f"Progress: {i}/{len(commits)} commits processed...")
+
+        pr_info = get_pr_for_commit(commit_sha, owner, repo, token)
+        if pr_info and pr_info["number"] not in seen_pr_numbers:
+            seen_pr_numbers.add(pr_info["number"])
+            prs.append(pr_info)
+
+    return prs
+
+
+def filter_prs(prs: List[Dict[str, Any]], ignore_patterns: List[str]) -> List[Dict[str, Any]]:
+    """Filter out PRs matching any of the ignore patterns."""
     compiled_patterns = [re.compile(pattern) for pattern in ignore_patterns]
 
     filtered = []
-    for message, author, sha in commits:
+    for pr in prs:
-        if not any(pattern.search(message) for pattern in compiled_patterns):
+        if not any(pattern.search(pr["title"]) for pattern in compiled_patterns):
-            filtered.append((message, author, sha))
+            filtered.append(pr)
 
     return filtered
 
 
-def generate_changelog(version: str, prev_version: str, commits: List[Tuple[str, str]], 
+def generate_changelog(version: str, prev_version: str, prs: List[Dict[str, Any]],
-                      cs_commit_new: Optional[str], cs_commit_old: Optional[str]) -> str:
+                      cs_commit_new: Optional[str], cs_commit_old: Optional[str],
+                      owner: str, repo: str) -> str:
     """Generate markdown changelog."""
     # Calculate statistics
-    commit_count = len(commits)
+    pr_count = len(prs)
-    unique_authors = len(set(author for _, author, _ in commits))
+    unique_authors = len(set(pr["author"] for pr in prs))
 
     changelog = f"# Dalamud Release v{version}\n\n"
     changelog += f"We just released Dalamud v{version}, which should be available to users within a few minutes. "
-    changelog += f"This release includes **{commit_count} commit{'s' if commit_count != 1 else ''} from {unique_authors} contributor{'s' if unique_authors != 1 else ''}**.\n"
+    changelog += f"This release includes **{pr_count} PR{'s' if pr_count != 1 else ''} from {unique_authors} contributor{'s' if unique_authors != 1 else ''}**.\n"
-    changelog += f"[Click here](<https://github.com/goatcorp/Dalamud/compare/{prev_version}...{version}>) to see all Dalamud changes.\n\n"
+    changelog += f"[Click here](<https://github.com/{owner}/{repo}/compare/{prev_version}...{version}>) to see all Dalamud changes.\n\n"
 
     if cs_commit_new and cs_commit_old and cs_commit_new != cs_commit_old:
         changelog += f"It ships with an updated **FFXIVClientStructs [`{cs_commit_new[:7]}`](<https://github.com/aers/FFXIVClientStructs/commit/{cs_commit_new}>)**.\n"
@@ -104,8 +191,8 @@ def generate_changelog(version: str, prev_version: str, commits: List[Tuple[str,
 
     changelog += "## Dalamud Changes\n\n"
 
-    for message, author, sha in commits:
+    for pr in prs:
-        changelog += f"* {message} (by **{author}** as [`{sha}`](<https://github.com/goatcorp/Dalamud/commit/{sha}>))\n"
+        changelog += f"* {pr['title']} ([#**{pr['number']}**](<{pr['url']}>) by **{pr['author']}**)\n"
 
     return changelog
 
@@ -158,11 +245,16 @@ def main():
         required=True,
         help="Discord webhook URL"
     )
+    parser.add_argument(
+        "--github-token",
+        default=os.environ.get("GITHUB_TOKEN"),
+        help="GitHub API token (or set GITHUB_TOKEN env var). Increases rate limit."
+    )
     parser.add_argument(
         "--ignore",
         action="append",
         default=[],
-        help="Regex patterns to ignore commits (can be specified multiple times)"
+        help="Regex patterns to ignore PRs (can be specified multiple times)"
     )
     parser.add_argument(
         "--submodule-path",
@@ -172,6 +264,10 @@ def main():
 
     args = parser.parse_args()
 
+    # Get repository info
+    owner, repo = get_repo_info()
+    print(f"Repository: {owner}/{repo}")
+
     # Get the last two tags
     latest_tag, previous_tag = get_last_two_tags()
     print(f"Generating changelog between {previous_tag} and {latest_tag}")
@@ -185,17 +281,18 @@ def main():
     if cs_commit_old:
         print(f"FFXIVClientStructs commit (old): {cs_commit_old[:7]}")
 
-    # Get commits between tags
+    # Get PRs between tags
-    commits = get_commits_between_tags(latest_tag, previous_tag)
+    prs = get_prs_between_tags(latest_tag, previous_tag, owner, repo, args.github_token)
-    print(f"Found {len(commits)} commits")
+    prs.reverse()
+    print(f"Found {len(prs)} PRs")
 
-    # Filter commits
+    # Filter PRs
-    filtered_commits = filter_commits(commits, args.ignore)
+    filtered_prs = filter_prs(prs, args.ignore)
-    print(f"After filtering: {len(filtered_commits)} commits")
+    print(f"After filtering: {len(filtered_prs)} PRs")
 
     # Generate changelog
-    changelog = generate_changelog(latest_tag, previous_tag, filtered_commits, 
+    changelog = generate_changelog(latest_tag, previous_tag, filtered_prs,
-                                   cs_commit_new, cs_commit_old)
+                                   cs_commit_new, cs_commit_old, owner, repo)
 
     print("\n" + "="*50)
     print("Generated Changelog:")

.github/workflows/generate-changelog.yml

@@ -6,6 +6,8 @@ on:
     tags:
       - '*'
 
+permissions: read-all
+
 jobs:
   generate-changelog:
     runs-on: ubuntu-latest
@@ -28,14 +30,14 @@ jobs:
           pip install requests
 
       - name: Generate and post changelog
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          GIT_TERMINAL_PROMPT: 0
         run: |
           python .github/generate_changelog.py \
             --webhook-url "${{ secrets.DISCORD_CHANGELOG_WEBHOOK_URL }}" \
-            --ignore "^Merge" \
+            --ignore "Update ClientStructs" \
-            --ignore "^build:" \
+            --ignore "^build:"
-            --ignore "^docs:"
-        env:
-          GIT_TERMINAL_PROMPT: 0
 
       - name: Upload changelog as artifact
         if: always()