Daily Trend Hunter #5
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: Daily Trend Hunter | |
| on: | |
| schedule: | |
| - cron: "0 8 * * *" # Runs at 8:00 AM UTC daily | |
| workflow_dispatch: | |
| jobs: | |
| hunt: | |
| name: Fetch Trends & Curate via AI | |
| runs-on: ubuntu-latest | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - name: Set up Python | |
| uses: actions/setup-python@v5 | |
| with: | |
| python-version: "3.11" | |
| - name: Install dependencies | |
| run: pip install requests pygithub litellm pydantic | |
| - name: Run Hunter & AI Curation Script | |
| env: | |
| GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} | |
| CF_ACCOUNT_ID: ${{ secrets.CLOUDFLARE_ACCOUNT_ID }} | |
| CF_GATEWAY_ID: ${{ secrets.CLOUDFLARE_GATEWAY_ID }} | |
| CF_API_TOKEN: ${{ secrets.CLOUDFLARE_AI_GATEWAY_TOKEN }} | |
| WORKER_API_URL: ${{ secrets.WORKER_API_URL_BASE }}/upsert/daily-trends | |
| WORKER_API_KEY: ${{ secrets.WORKER_API_KEY }} | |
| run: | | |
| cat << 'EOF' > hunter.py | |
| import os | |
| import requests | |
| import json | |
| from datetime import datetime, timedelta | |
| from github import Github | |
| from pydantic import BaseModel | |
| from litellm import completion | |
| # --- 1. Define the Structured Output Schema --- | |
| class CuratedRepo(BaseModel): | |
| name: str | |
| url: str | |
| category: str | |
| why_its_interesting: str | |
| innovation_score: int | |
| class DailyReport(BaseModel): | |
| trend_summary: str | |
| top_picks: list[CuratedRepo] | |
| def get_raw_trends(g): | |
| # Get User's Existing Stars (to avoid duplicates) | |
| # Note: PyGithub get_starred() can be slow if you have thousands of stars. | |
| me = g.get_user("${{ github.repository_owner }}") | |
| starred_ids = set() | |
| print("Fetching user stars...") | |
| for repo in me.get_starred(): | |
| starred_ids.add(repo.id) | |
| date_since = (datetime.now() - timedelta(days=30)).strftime('%Y-%m-%d') | |
| queries = [ | |
| f"language:typescript created:>{date_since} sort:stars", | |
| f"language:python created:>{date_since} sort:stars", | |
| f"topic:google-apps-script created:>{date_since}", | |
| f"topic:cloudflare-workers created:>{date_since}" | |
| ] | |
| watched_orgs = ["cloudflare", "openai", "google-gemini", "langchain-ai"] | |
| raw_repos = [] | |
| print("Scanning for trending repos...") | |
| for q in queries: | |
| repos = g.search_repositories(query=q) | |
| count = 0 | |
| for repo in repos: | |
| if count >= 10: break # Top 10 per category | |
| if repo.id in starred_ids: continue | |
| raw_repos.append({ | |
| "name": repo.full_name, | |
| "description": repo.description or "No description", | |
| "url": repo.html_url, | |
| "category": q.split(" ")[0] | |
| }) | |
| count += 1 | |
| print("Scanning watched orgs...") | |
| for org_name in watched_orgs: | |
| try: | |
| org = g.get_organization(org_name) | |
| for repo in org.get_repos(sort="created", direction="desc"): | |
| if (datetime.now() - repo.created_at).days > 30: break | |
| raw_repos.append({ | |
| "name": repo.full_name, | |
| "description": repo.description or "No description", | |
| "url": repo.html_url, | |
| "category": "org_watch" | |
| }) | |
| except Exception as e: | |
| print(f"Skipping {org_name}: {e}") | |
| return raw_repos | |
| def curate_with_ai(raw_repos): | |
| print(f"Sending {len(raw_repos)} repos to Cloudflare AI Gateway for curation...") | |
| api_base = f"https://gateway.ai.cloudflare.com/v1/{os.environ['CF_ACCOUNT_ID']}/{os.environ['CF_GATEWAY_ID']}/workers-ai/v1" | |
| prompt = f"Review the following list of newly created repositories. Select the top 5 most innovative or interesting ones, and provide a brief summary of today's trends.\n\nData: {json.dumps(raw_repos)}" | |
| response = completion( | |
| model="openai/@cf/openai/gpt-oss-120b", | |
| api_base=api_base, | |
| api_key=os.environ["CF_API_TOKEN"], | |
| messages=[ | |
| {"role": "system", "content": "You are an expert developer curating daily tech trends. Return ONLY valid JSON matching the requested schema."}, | |
| {"role": "user", "content": prompt} | |
| ], | |
| response_format=DailyReport, | |
| max_tokens=4096, | |
| extra_body={"reasoning": {"effort": "high"}} | |
| ) | |
| return response.choices[0].message.content | |
| def main(): | |
| g = Github(os.environ["GITHUB_TOKEN"]) | |
| # 1. Fetch raw data from GitHub | |
| raw_repos = get_raw_trends(g) | |
| # 2. Curate using Cloudflare Workers AI + LiteLLM | |
| ai_curated_json_str = curate_with_ai(raw_repos) | |
| # 3. Post to Worker | |
| worker_url = os.environ.get("WORKER_API_URL") | |
| if worker_url: | |
| print("Posting curated trends to Worker API...") | |
| # Parse string to dict to ensure valid JSON payload | |
| payload = json.loads(ai_curated_json_str) | |
| payload["date"] = datetime.now().isoformat() | |
| headers = { | |
| "Content-Type": "application/json", | |
| "X-API-Key": os.environ["WORKER_API_KEY"], | |
| "User-Agent": "github-actions/hunter" | |
| } | |
| resp = requests.post(worker_url, json=payload, headers=headers) | |
| resp.raise_for_status() | |
| print("Successfully uploaded to Worker.") | |
| if __name__ == "__main__": | |
| main() | |
| EOF | |
| python hunter.py |