#!/usr/bin/env python3 """ Post a welcome comment on a pull request labeled `first contribution`, unless the community bot has already welcomed its author on any of their pull requests. Requires: requests (pip install requests) Usage: python github-welcome-first-contribution.py [--dry-run] Environment variables: GITHUB_TOKEN - GitHub token (issues: write, pull requests: read) """ import argparse from github_helpers import ( github_rest_api, github_rest_get_paginated, post_github_comment, ) REPO_OWNER = "zed-industries" REPO_NAME = "zed" FIRST_CONTRIBUTION_LABEL = "first contribution" COMMUNITY_BOT_LOGIN = "zed-community-bot[bot]" WELCOME_MARKER = "" SEARCH_PAGE_SIZE = 100 # The search API returns at most 1000 results which should never be relevant here but hey. MAX_SEARCH_PAGES = 10 WELCOME_COMMENT = f"""{WELCOME_MARKER} Thanks for your first contribution to Zed! For the best chances of merge, please read our [contributing guide](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md) if you haven't already. The key points for a first PR: - **Show how you tested it.** Add evidence that fits the change: screenshots/video for UI changes, benchmarks for performance work, etc. Using LLMs or other help is fine, but you must understand and test the changes yourself. **Write review replies in your own words;** a human is reading them. - **One PR at a time.** Things go smoother once your first PR lands; several open at once tend to go stale and collect the same feedback. - **Discuss features first.** If this adds a feature that staff haven't agreed to in an issue, please start a [GitHub discussion](https://github.com/zed-industries/zed/discussions) instead of a PR. - **Don't ping staff by username** unless they've told you that's okay.""" def search_pull_requests_by_author(author): pull_requests = [] for page in range(1, MAX_SEARCH_PAGES + 1): result = github_rest_api( "GET", "search/issues", params={ "q": f"repo:{REPO_OWNER}/{REPO_NAME} is:pr author:{author}", "sort": "created", "order": "asc", "per_page": SEARCH_PAGE_SIZE, "page": page, }, ) pull_requests.extend(result["items"]) if len(result["items"]) < SEARCH_PAGE_SIZE: break return pull_requests def find_welcomed_pull_request(pull_requests): for pull_request in pull_requests: comments = github_rest_get_paginated( f"repos/{REPO_OWNER}/{REPO_NAME}/issues/{pull_request['number']}/comments" ) if any( (comment["user"] or {}).get("login") == COMMUNITY_BOT_LOGIN and WELCOME_MARKER in (comment["body"] or "") for comment in comments ): return pull_request return None def welcome_first_time_contributor(pr_number, dry_run): pull_request = github_rest_api( "GET", f"repos/{REPO_OWNER}/{REPO_NAME}/pulls/{pr_number}" ) author = pull_request["user"]["login"] if not any( label["name"] == FIRST_CONTRIBUTION_LABEL for label in pull_request["labels"] ): print(f"PR #{pr_number} has no '{FIRST_CONTRIBUTION_LABEL}' label") return # Search results can lag behind a just-opened PR, so check it explicitly. welcomed_pull_request = find_welcomed_pull_request([pull_request]) if welcomed_pull_request is None: other_pull_requests = [ other_pull_request for other_pull_request in search_pull_requests_by_author(author) if other_pull_request["number"] != pr_number ] welcomed_pull_request = find_welcomed_pull_request(other_pull_requests) if welcomed_pull_request is not None: print( f"{author} was already welcomed on PR #{welcomed_pull_request['number']}" ) return if dry_run: print(f"Would post on PR #{pr_number}:\n\n{WELCOME_COMMENT}") else: post_github_comment(REPO_OWNER, REPO_NAME, pr_number, WELCOME_COMMENT) if __name__ == "__main__": argument_parser = argparse.ArgumentParser() argument_parser.add_argument("pr_number", type=int) argument_parser.add_argument( "--dry-run", action="store_true", help="Print the comment instead of posting it", ) arguments = argument_parser.parse_args() welcome_first_time_contributor(arguments.pr_number, arguments.dry_run)