#!/usr/bin/env python3 """ github-radar.py - Catch viral repos before they go mainstream Usage: github-radar.py # Full radar sweep github-radar.py --days 3 # Shorter lookback (default: 7) github-radar.py --save # Save to file github-radar.py --hn-only # Only Hacker News """ import json import sys import os import urllib.request import urllib.error import urllib.parse from datetime import datetime, timezone, timedelta import re import ssl CONFIG_FILE = os.path.join(os.environ.get("DIGEST_HOME", os.path.expanduser("~/.config/github-digest")), "github-radar.json") OUTPUT_FILE = os.path.join(os.environ.get("DIGEST_HOME", os.path.expanduser("~/.config/github-digest")), "github-radar.md") GITHUB_API = "https://api.github.com" HN_API = "https://hn.algolia.com/api/v1/search_by_date" def api_request(url, headers=None): """Generic API request.""" req = urllib.request.Request(url) req.add_header("User-Agent", "github-radar-script") if headers: for k, v in headers.items(): req.add_header(k, v) token = os.environ.get("GITHUB_TOKEN") if token and "github.com" in url: req.add_header("Authorization", f"token {token}") try: ctx = ssl.create_default_context() ctx.check_hostname = False ctx.verify_mode = ssl.CERT_NONE with urllib.request.urlopen(req, timeout=15, context=ctx) as resp: return json.loads(resp.read().decode()) except urllib.error.HTTPError as e: if e.code == 403: return {"error": "Rate limited"} return {"error": f"HTTP {e.code}"} except Exception as e: return {"error": str(e)} def parse_github_time(ts): if not ts: return None try: return datetime.fromisoformat(ts.replace("Z", "+00:00")) except: return None def days_since(dt): if not dt: return 999 return (datetime.now(timezone.utc) - dt).days def compute_velocity(stars, created_at): """Stars per day since creation.""" dt = parse_github_time(created_at) days = max(days_since(dt), 1) return round(stars / days, 1) def score_relevance(text, keywords): """Score how relevant text is to keywords.""" text_lower = text.lower() score = 0 for kw in keywords: if kw.lower() in text_lower: score += 10 # Bonus for being in name/title if kw.lower() in text_lower.split()[:5]: score += 20 return score def should_exclude(repo, patterns): """Check if repo matches exclusion patterns.""" text = f"{repo.get('name', '')} {repo.get('description', '')}".lower() for pattern in patterns: if pattern.lower() in text: return True return False def search_github(query, min_stars=0): """Search GitHub repos.""" # GitHub search API: encode query properly # Spaces in query should be +, but we pass to q= parameter encoded = urllib.parse.quote_plus(query) url = f"{GITHUB_API}/search/repositories?q={encoded}&sort=stars&order=desc&per_page=30" data = api_request(url) if not data or "error" in data: return [] return data.get("items", []) def scan_hackernews(keywords, days=7): """Scan HN Show HN for relevant posts.""" cutoff = datetime.now(timezone.utc) - timedelta(days=days) # Search Show HN posts from last N days hits = [] page = 0 while page < 3: # Max 3 pages url = f"{HN_API}?tags=show_hn&numericFilters=created_at_i>{int(cutoff.timestamp())}&hitsPerPage=30&page={page}" data = api_request(url) if not data or "hits" not in data: break for hit in data["hits"]: title = hit.get("title", "") text = hit.get("text", "") combined = f"{title} {text}" relevance = score_relevance(combined, keywords) if relevance >= 20: # At least 2 keyword matches hits.append({ "title": title, "url": hit.get("url", f"https://news.ycombinator.com/item?id={hit.get('objectID')}"), "points": hit.get("points", 0), "comments": hit.get("num_comments", 0), "relevance": relevance, "author": hit.get("author", "?"), }) if len(data["hits"]) < 30: break page += 1 # Sort by relevance then points hits.sort(key=lambda x: (x["relevance"], x["points"]), reverse=True) return hits[:10] def generate_radar(config, days=7): """Generate radar report.""" lines = [] date_format = "%Y-%m-%d" since_date = (datetime.now(timezone.utc) - timedelta(days=days)).strftime(date_format) lines.append("# 🎯 GitHub Trending Radar") lines.append(f"*{datetime.now().strftime('%Y-%m-%d %H:%M')} | {days} day lookback | Catching viral before mainstream*") lines.append("") lines.append("---") lines.append("") filters = config.get("filters", {}) exclude_patterns = filters.get("exclude_patterns", []) min_velocity = filters.get("min_stars_velocity_per_day", 20) all_findings = [] # GitHub searches for search in config.get("github_search_queries", []): query_template = search["query"] query = query_template % since_date keywords = search.get("relevance_keywords", []) min_stars = search.get("min_stars", 100) repos = search_github(query, min_stars) for repo in repos: if filters.get("exclude_forks", True) and repo.get("fork", False): continue if filters.get("exclude_archived", True) and repo.get("archived", False): continue if should_exclude(repo, exclude_patterns): continue stars = repo.get("stargazers_count", 0) created = repo.get("created_at", "") pushed = repo.get("pushed_at", "") velocity = compute_velocity(stars, created) if velocity < min_velocity: continue age_days = days_since(parse_github_time(created)) if age_days > filters.get("max_age_days", 30): continue # Relevance score desc = repo.get("description", "") or "" relevance = score_relevance(f"{repo['name']} {desc}", keywords) # Total score total_score = relevance + (velocity * 2) + (stars // 100) if total_score < config.get("output", {}).get("min_total_score", 50): continue all_findings.append({ "source": "github", "category": search["name"], "name": repo["full_name"], "url": repo["html_url"], "description": desc[:150] + "..." if len(desc) > 150 else desc, "stars": stars, "velocity": velocity, "age_days": age_days, "relevance": relevance, "score": total_score, "language": repo.get("language", "?"), }) # Deduplicate by name seen = set() unique = [] for f in sorted(all_findings, key=lambda x: x["score"], reverse=True): if f["name"] not in seen: seen.add(f["name"]) unique.append(f) # Take top N max_results = config.get("output", {}).get("max_results_per_query", 10) github_findings = unique[:max_results * 2] # Hacker News hn_keywords = [] for search in config.get("github_search_queries", []): hn_keywords.extend(search.get("relevance_keywords", [])) hn_keywords = list(set(hn_keywords)) hn_findings = scan_hackernews(hn_keywords, days) # Output GitHub findings if github_findings: lines.append("## 🔥 GitHub Trending") lines.append(f"*Repos gaining stars fast | Velocity > {min_velocity}/day*") lines.append("") for i, f in enumerate(github_findings, 1): highlight = "🔴" if f["score"] >= config.get("output", {}).get("highlight_threshold", 500) else "" lines.append(f"### {i}. [{f['name']}]({f['url']}) {highlight}") lines.append(f"*{f['description']}*") lines.append("") lines.append(f"| Stars | Velocity | Age | Language | Category |") lines.append(f"|-------|----------|-----|----------|----------|") lines.append(f"| {f['stars']:,} | {f['velocity']}/day | {f['age_days']}d | {f['language']} | {f['category']} |") lines.append("") # Why it matters if f["velocity"] > 100: lines.append(f"⚡ **Viral**: {f['velocity']} stars/day — mainstream attention imminent") elif "agent" in f["name"].lower() or "agent" in f["description"].lower(): lines.append(f"🤖 **Agent**: Related to AI agents — check if it replaces/supplements your stack") elif "claude" in f["description"].lower() or "cursor" in f["description"].lower(): lines.append(f"💻 **Coding Tool**: Could be a Cursor/Claude alternative — worth evaluating") elif "local" in f["description"].lower(): lines.append(f"🏠 **Local**: Self-hosted/privacy focused — aligns with your Mac Mini setup") else: lines.append(f"📈 **Trending**: Fast star growth in {f['category']} category") lines.append("") else: lines.append("## 🔥 GitHub Trending") lines.append("*Nothing viral in the last {} days. Quiet period, or the algo missed it.*".format(days)) lines.append("") # Output HN findings if hn_findings: lines.append("## 📰 Hacker News - Show HN") lines.append("*Early adopter projects. Often precede GitHub virality by 1-3 days.*") lines.append("") for i, f in enumerate(hn_findings[:5], 1): lines.append(f"### {i}. [{f['title']}]({f['url']})") lines.append(f"👍 {f['points']} points | 💬 {f['comments']} comments | by {f['author']} | Relevance: {f['relevance']}") lines.append("") lines.append("---") lines.append(f"*Scanned {len(config.get('github_search_queries', []))} categories | {len(github_findings)} GitHub hits | {len(hn_findings)} HN hits*") lines.append("") lines.append("💡 **How to use this:**") lines.append("- 🔴 = Exploding fast, check NOW") lines.append("- High velocity (>100/day) = Will be everywhere in 48h") lines.append("- HN posts with 100+ points = Early adopter validation") lines.append("- Click through, read README, decide if it fits your stack") lines.append("") lines.append("📋 *Edit repos/categories: `$DIGEST_HOME/github-radar.json`*") return "\n".join(lines) def main(): days = 7 save_mode = False hn_only = False args = sys.argv[1:] i = 0 while i < len(args): if args[i] == "--days" and i + 1 < len(args): days = int(args[i + 1]) i += 2 elif args[i] == "--save": save_mode = True i += 1 elif args[i] == "--hn-only": hn_only = True i += 1 else: i += 1 try: with open(CONFIG_FILE) as f: config = json.load(f) except FileNotFoundError: print(f"Config not found: {CONFIG_FILE}") sys.exit(1) print(f"🎯 Scanning for viral repos (last {days} days)...", file=sys.stderr) report = generate_radar(config, days) if save_mode: os.makedirs(os.path.dirname(OUTPUT_FILE), exist_ok=True) with open(OUTPUT_FILE, "w") as f: f.write(report) print(f"Saved to: {OUTPUT_FILE}") else: print(report) if __name__ == "__main__": main()