#!/usr/bin/env python3 """ arxiv_daily.py - Daily arXiv scanner for Bo Wang Fetches top 5 papers from AI, biology, medicine categories (abstracts only, no PDFs) """ import urllib.request import urllib.parse import xml.etree.ElementTree as ET import re import sys from datetime import datetime, timedelta, timezone # ── Interest scoring ────────────────────────────────────────────────────────── HIGH_PRIORITY = [ r"single.cell", r"scrna.?seq", r"scatac", r"spatial transcriptomics", r"cell atlas", r"drug discovery", r"drug design", r"molecular generation", r"protein structure", r"foundation model", r"virtual cell", r"perturbation", r"gene regulatory network", r"cell type annotation", r"multiom", r"multi.om", r"scgpt", r"single.cell foundation", r"cancer genomics", r"epigenomics", r"crispr", r"rna.seq", r"trajectory inference", ] MEDIUM_PRIORITY = [ r"large language model", r"\bllm\b", r"transformer", r"reinforcement learning", r"\bagent\b", r"multi.agent", r"generative model", r"diffusion model", r"genomics", r"proteomics", r"biomarker", r"clinical", r"medical imaging", r"federated learning", r"zero.shot", r"few.shot", r"self.supervised", r"representation learning", r"neural network", r"deep learning", r"drug target", r"molecular docking", r"sequence model", ] CATEGORIES = [ "cs.LG", "cs.AI", "cs.CL", "cs.NE", "q-bio.GN", "q-bio.QM", "q-bio.BM", "q-bio.MN", "q-bio.CB", "stat.ML", ] # ── Fetch ───────────────────────────────────────────────────────────────────── def fetch_papers_huggingface(max_results=100): """Fetch trending arXiv papers from HuggingFace Daily Papers API.""" import json from datetime import date, timedelta as td papers = [] seen = set() # Try today and yesterday for delta in range(3): day = (date.today() - td(days=delta)).strftime("%Y-%m-%d") url = f"https://huggingface.co/api/daily_papers?date={day}" req = urllib.request.Request(url, headers={"User-Agent": "arxiv-daily-bot/1.0"}) try: with urllib.request.urlopen(req, timeout=15) as resp: items = json.loads(resp.read().decode("utf-8")) for item in items: p = item.get("paper", {}) pid = p.get("id", "") if not pid or pid in seen: continue seen.add(pid) authors_raw = p.get("authors") or [] authors = [a.get("name", "") for a in authors_raw[:4]] papers.append({ "title": (p.get("title") or "").strip(), "abstract": (p.get("summary") or "").strip(), "url": f"https://arxiv.org/abs/{pid}", "published": day + "T00:00:00+00:00", "authors": authors, "categories": [], "upvotes": p.get("upvotes", 0), }) except Exception: pass return papers def fetch_papers(max_results=100): """Fetch papers — HuggingFace Daily Papers as primary source.""" papers = fetch_papers_huggingface(max_results) if papers: return papers return None # ── Parse ───────────────────────────────────────────────────────────────────── def parse_papers(xml_data): ns_atom = "http://www.w3.org/2005/Atom" ns_arxiv = "http://arxiv.org/schemas/atom" root = ET.fromstring(xml_data) papers = [] for entry in root.findall(f"{{{ns_atom}}}entry"): def t(tag, ns=ns_atom): el = entry.find(f"{{{ns}}}{tag}") return el.text.strip() if el is not None and el.text else "" title = re.sub(r"\s+", " ", t("title")) abstract = re.sub(r"\s+", " ", t("summary")) url = t("id") published= t("published") authors = [ a.find(f"{{{ns_atom}}}name").text for a in entry.findall(f"{{{ns_atom}}}author") if a.find(f"{{{ns_atom}}}name") is not None ] # categories cats = set() pc = entry.find(f"{{{ns_arxiv}}}primary_category") if pc is not None: cats.add(pc.get("term", "")) for c in entry.findall(f"{{{ns_atom}}}category"): cats.add(c.get("term", "")) papers.append({ "title": title, "abstract": abstract, "url": url.replace("http://", "https://"), "published": published, "authors": authors[:4], "categories": sorted(cats - {""}), }) return papers # ── Score ───────────────────────────────────────────────────────────────────── def score(paper): text = (paper["title"] + " " + paper["abstract"]).lower() s = 0 for kw in HIGH_PRIORITY: if re.search(kw, text): s += 3 for kw in MEDIUM_PRIORITY: if re.search(kw, text): s += 1 return s # ── Format ──────────────────────────────────────────────────────────────────── def format_report(papers, mode="telegram"): today = datetime.now().strftime("%B %d, %Y") if mode == "telegram": lines = [f"📚 *arXiv Daily Scout — {today}*\n"] for i, p in enumerate(papers, 1): authors = ", ".join(p["authors"][:2]) if len(p["authors"]) > 2: authors += " et al." snippet = p["abstract"][:280].rstrip() + "…" cats = " ".join(f"`{c}`" for c in p["categories"][:3]) lines.append( f"*{i}. {p['title']}*\n" f"_{authors}_ {cats}\n" f"{snippet}\n" f"🔗 {p['url']}\n" ) return "\n".join(lines) else: # email / plain lines = [f"arXiv Daily Scout — {today}\n{'='*50}\n"] for i, p in enumerate(papers, 1): authors = ", ".join(p["authors"][:2]) if len(p["authors"]) > 2: authors += " et al." snippet = p["abstract"][:400].rstrip() + "…" cats = ", ".join(p["categories"][:3]) lines.append( f"{i}. {p['title']}\n" f" {authors} [{cats}]\n" f" {snippet}\n" f" {p['url']}\n" ) return "\n".join(lines) # ── Main ────────────────────────────────────────────────────────────────────── if __name__ == "__main__": mode = sys.argv[1] if len(sys.argv) > 1 else "telegram" result = fetch_papers(max_results=100) if not result: print("⚠️ All paper sources failed or returned no results.", file=sys.stderr) sys.exit(1) papers = result # Score for p in papers: p["score"] = score(p) # Prefer papers submitted in the last 2 days; fall back to all if sparse cutoff = datetime.now(timezone.utc) - timedelta(days=2) recent = [ p for p in papers if datetime.fromisoformat(p["published"].replace("Z", "+00:00")) > cutoff ] pool = recent if len(recent) >= 5 else papers top5 = sorted(pool, key=lambda x: (x["score"], x.get("upvotes", 0)), reverse=True)[:5] print(format_report(top5, mode=mode))