action-release-notes/scripts/generate_notes.py

#!/usr/bin/env python3
"""
generate_notes.py - Generate Markdown release notes from git commits using OmniRoute or OpenAI-compatible endpoints.
Works standalone or as part of a Gitea / GitHub composite action.
"""

import argparse
import json
import os
import re
import subprocess
import sys
import urllib.error
import urllib.request
import uuid


def parse_args():
    parser = argparse.ArgumentParser(description="Generate AI release notes from git commits.")
    parser.add_argument("--api-key", default=os.environ.get("INPUT_API_KEY", os.environ.get("OMNI_API_KEY", "")))
    parser.add_argument("--base-url", default=os.environ.get("INPUT_BASE_URL", "https://omni.freakma.com/v1"))
    parser.add_argument("--model", default=os.environ.get("INPUT_MODEL", "auto/fast"))
    parser.add_argument("--version", default=os.environ.get("INPUT_VERSION", ""))
    parser.add_argument("--from-ref", default=os.environ.get("INPUT_FROM_REF", ""))
    parser.add_argument("--to-ref", default=os.environ.get("INPUT_TO_REF", "HEAD"))
    parser.add_argument("--tag-pattern", default=os.environ.get("INPUT_TAG_PATTERN", "v*"))
    parser.add_argument("--max-commits", type=int, default=int(os.environ.get("INPUT_MAX_COMMITS", "50")))
    parser.add_argument("--fallback-notes", default=os.environ.get("INPUT_FALLBACK_NOTES", "Maintenance updates and performance improvements."))
    parser.add_argument("--timeout", type=int, default=int(os.environ.get("INPUT_TIMEOUT", "25")))
    parser.add_argument("--temperature", type=float, default=float(os.environ.get("INPUT_TEMPERATURE", "0.2")))
    parser.add_argument("--system-prompt", default=os.environ.get("INPUT_SYSTEM_PROMPT", ""))
    parser.add_argument("--save-to-file", default=os.environ.get("INPUT_SAVE_TO_FILE", ""))
    return parser.parse_args()


def run_git(args):
    try:
        res = subprocess.run(["git"] + args, capture_output=True, text=True, check=False)
        return res.returncode, res.stdout.strip(), res.stderr.strip()
    except Exception as e:
        return 1, "", str(e)


def find_previous_tag(to_ref, tag_pattern, version=""):
    # 0. If a target version is provided (e.g. 1.8.0), check if tag v$VERSION already exists.
    # If it does, the commits for this release start from the tag BEFORE v$VERSION:
    if version:
        target_tag = version if version.startswith("v") else f"v{version}"
        code_prev, prev_tag, _ = run_git(["describe", "--tags", f"--match={tag_pattern}", "--abbrev=0", f"{target_tag}^"])
        if code_prev == 0 and prev_tag:
            return prev_tag

    # 1. Check if to_ref itself is tagged with an exact tag matching tag_pattern:
    if tag_pattern:
        code, stdout, _ = run_git(["describe", "--tags", f"--match={tag_pattern}", "--exact-match", to_ref])
        if code == 0 and stdout:
            # to_ref is an existing release tag (e.g. tag was just created on HEAD).
            # The prior tag must come from to_ref's parent commit.
            code_prev, prev_tag, _ = run_git(["describe", "--tags", f"--match={tag_pattern}", "--abbrev=0", f"{to_ref}^"])
            if code_prev == 0 and prev_tag:
                return prev_tag

    # 2. to_ref is not an exact tag; find the most recent tag matching pattern reaching to_ref:
    if tag_pattern:
        code, stdout, _ = run_git(["describe", "--tags", f"--match={tag_pattern}", "--abbrev=0", to_ref])
        if code == 0 and stdout:
            return stdout

    # 3. Fallback: try describe without tag pattern:
    code, stdout, _ = run_git(["describe", "--tags", "--exact-match", to_ref])
    if code == 0 and stdout:
        code_prev, prev_tag, _ = run_git(["describe", "--tags", "--abbrev=0", f"{to_ref}^"])
        if code_prev == 0 and prev_tag:
            return prev_tag

    code, stdout, _ = run_git(["describe", "--tags", "--abbrev=0", to_ref])
    if code == 0 and stdout:
        return stdout

    return None


def collect_commits(from_ref, to_ref, tag_pattern, max_commits, version=""):
    # Ensure tags are up-to-date
    run_git(["fetch", "--tags", "--prune"])

    if not from_ref:
        from_ref = find_previous_tag(to_ref, tag_pattern, version)

    commits = ""
    if from_ref:
        sys.stderr.write(f"[action-release-notes] Comparing commits from {from_ref} to {to_ref}...\n")
        code, stdout, _ = run_git(["log", f"{from_ref}..{to_ref}", "--pretty=format:- %s (%h)"])
        if code == 0 and stdout:
            commits = stdout

    if not commits:
        sys.stderr.write(f"[action-release-notes] No range commits found or no prior tag; falling back to last {max_commits} commits on {to_ref}...\n")
        code, stdout, _ = run_git(["log", f"-n{max_commits}", "--pretty=format:- %s (%h)", to_ref])
        if code == 0 and stdout:
            commits = stdout

    if not commits:
        return ""

    # Filter out noisy commits (version bump, chores, merge pull requests)
    filtered = []
    for line in commits.splitlines():
        line_clean = line.strip()
        if not line_clean:
            continue
        # Drop version bumps
        if re.search(r'chore(\(ci\))?:\s*(bump version|release)', line_clean, re.I):
            continue
        filtered.append(line_clean)

    return "\n".join(filtered) if filtered else commits


def clean_llm_markdown(content):
    # Remove thinking tags from reasoning models (e.g. Qwen, DeepSeek)
    content = re.sub(r'<think>.*?</think>', '', content, flags=re.DOTALL).strip()
    
    # Strip markdown code fences if model wrapped entire answer
    lines = content.splitlines()
    if len(lines) >= 2:
        if lines[0].strip().startswith("```"):
            lines = lines[1:]
        if lines and lines[-1].strip().startswith("```"):
            lines = lines[:-1]
    
    return "\n".join(lines).strip()


def query_llm(args, commits_text):
    if not args.api_key:
        sys.stderr.write("[action-release-notes] Warning: No API key provided; using fallback release notes.\n")
        return args.fallback_notes

    default_system = (
        "You are an automated release notes generator. Given git commit messages, "
        "produce concise, structured Markdown release notes grouped into categories "
        "like '### Features', '### Bug Fixes', and '### Improvements'. "
        "Filter out chores and noise. Output ONLY valid Markdown release notes without "
        "conversational preamble, thinking blocks, or code fences."
    )
    system_prompt = args.system_prompt if args.system_prompt else default_system

    version_str = f" v{args.version}" if args.version else ""
    user_prompt = f"Generate concise release notes for{version_str} based on the following git commits:\n\n{commits_text}"

    payload = {
        "model": args.model,
        "messages": [
            {"role": "system", "content": system_prompt},
            {"role": "user", "content": user_prompt}
        ],
        "temperature": args.temperature
    }

    url = f"{args.base_url.rstrip('/')}/chat/completions"
    req = urllib.request.Request(
        url,
        data=json.dumps(payload).encode("utf-8"),
        headers={
            "Authorization": f"Bearer {args.api_key}",
            "Content-Type": "application/json",
            "User-Agent": "gitea-action-release-notes"
        }
    )

    try:
        sys.stderr.write(f"[action-release-notes] Requesting release notes from {url} using model {args.model}...\n")
        with urllib.request.urlopen(req, timeout=args.timeout) as response:
            resp_bytes = response.read()
            data = json.loads(resp_bytes.decode("utf-8"))
            content = data["choices"][0]["message"]["content"]
            cleaned = clean_llm_markdown(content)
            if cleaned:
                return cleaned
            sys.stderr.write("[action-release-notes] Model returned empty content; using fallback.\n")
            return args.fallback_notes
    except Exception as e:
        sys.stderr.write(f"[action-release-notes] LLM query failed ({e}); using fallback release notes.\n")
        return args.fallback_notes


def main():
    args = parse_args()
    commits = collect_commits(args.from_ref, args.to_ref, args.tag_pattern, args.max_commits, args.version)
    
    if not commits:
        sys.stderr.write("[action-release-notes] No commits discovered; using fallback release notes.\n")
        notes = args.fallback_notes
    else:
        notes = query_llm(args, commits)

    # Save to file if requested
    notes_file_path = ""
    if args.save_to_file:
        try:
            with open(args.save_to_file, "w", encoding="utf-8") as f:
                f.write(notes + "\n")
            notes_file_path = os.path.abspath(args.save_to_file)
            sys.stderr.write(f"[action-release-notes] Notes saved to {notes_file_path}\n")
        except Exception as e:
            sys.stderr.write(f"[action-release-notes] Failed to save notes to file: {e}\n")

    # Write to GITHUB_OUTPUT if available
    gh_output = os.environ.get("GITHUB_OUTPUT")
    if gh_output and os.path.exists(os.path.dirname(gh_output) or "."):
        delim = f"EOF_{uuid.uuid4().hex}"
        try:
            with open(gh_output, "a", encoding="utf-8") as f:
                f.write(f"notes<<{delim}\n")
                f.write(f"{notes}\n")
                f.write(f"{delim}\n")
                f.write(f"notes_file={notes_file_path}\n")
        except Exception as e:
            sys.stderr.write(f"[action-release-notes] Failed to write to GITHUB_OUTPUT: {e}\n")

    # Output notes to stdout for build log visibility
    print(notes)


if __name__ == "__main__":
    main()