Update mcp/arxiv.py

mcp/arxiv.py

@@ -1,84 +1,20 @@
-#!/usr/bin/env python3
-"""
-MedGenesis – arXiv async fetcher (Atom API).
-
-* Uses HTTPS (`https://export.arxiv.org/...`) to avoid HTTP 301 redirects.
-* Async httpx fetch with 2×/4× exponential-back-off retry.
-* Parses the Atom feed with feedparser inside a thread (non-blocking).
-* 6-hour LRU cache keyed by “query+max_results”.
-* Returns a list of dicts matching schemas.Paper.
-
-API docs: https://arxiv.org/help/api/user-manual
-"""
-from __future__ import annotations
-
-import asyncio
-from functools import lru_cache
-from typing import List, Dict
-from urllib.parse import quote_plus
-
+# mcp/arxiv.py
 import feedparser
-import httpx
-
-_BASE   = "https://export.arxiv.org/api/query?search_query="
-_TIMEOUT = 10
-_MAX_RES = 25
-_HEADERS = {"User-Agent": "MedGenesis/1.0 (https://huggingface.co/spaces)"}
-
-
-# ──────────────────────────────────────────────────────────────────────
-# Internal fetch helper with retry
-# ──────────────────────────────────────────────────────────────────────
-async def _fetch_raw(query: str, max_results: int, *, retries: int = 3) -> str:
-    """Return raw Atom XML from arXiv."""
-    max_results = max(1, min(max_results, _MAX_RES))
-    url = f"{_BASE}{quote_plus(query)}&max_results={max_results}"
-    delay = 2
-    last: httpx.Response | None = None
-    for _ in range(retries):
-        async with httpx.AsyncClient(timeout=_TIMEOUT, headers=_HEADERS) as cli:
-            last = await cli.get(url)
-            if last.status_code == 200:
-                return last.text
-        await asyncio.sleep(delay)
-        delay *= 2
-    raise RuntimeError(f"arXiv API failed: {last.status_code if last else 'No response'}")
-
-
-# ──────────────────────────────────────────────────────────────────────
-# Public cached fetch + parse
-# ──────────────────────────────────────────────────────────────────────
-@lru_cache(maxsize=256)
-async def fetch_arxiv(query: str, *, max_results: int = 5) -> List[Dict]:
-    """Return arXiv paper dicts compatible with schemas.Paper."""
-    xml_text = await _fetch_raw(query, max_results)
-
-    # feedparser is blocking; parse in thread
-    feed = await asyncio.to_thread(feedparser.parse, xml_text)
+from urllib.parse import quote_plus
 
-    papers: List[Dict] = []
-    for ent in feed.entries:
-        authors = (
-            ", ".join(a.name for a in getattr(ent, "authors", []))
-            if hasattr(ent, "authors") else "Unknown"
-        )
-        papers.append({
-            "title"    : getattr(ent, "title", "[No title]"),
-            "authors"  : authors,
-            "summary"  : getattr(ent, "summary", ""),
-            "link"     : getattr(ent, "link", ""),
-            "published": getattr(ent, "published", ""),
-            "source"   : "arXiv",
+ARXIV_BASE = "http://export.arxiv.org/api/query?search_query="
+
+async def fetch_arxiv(query: str, max_results: int = 5) -> list[dict]:
+    url = f"{ARXIV_BASE}{quote_plus(query)}&max_results={max_results}"
+    feed = feedparser.parse(url)
+    out = []
+    for e in feed.entries:
+        out.append({
+            "title": e.get("title",""),
+            "authors": ", ".join([a.name for a in getattr(e,"authors",[])]),
+            "summary": e.get("summary",""),
+            "link": e.get("link",""),
+            "published": getattr(e,"published",""),
+            "source": "arXiv"
         })
-    return papers
-
-
-# ──────────────────────────────────────────────────────────────────────
-# CLI demo
-# ──────────────────────────────────────────────────────────────────────
-if __name__ == "__main__":
-    async def _demo():
-        papers = await fetch_arxiv("glioblastoma CRISPR", max_results=3)
-        for p in papers:
-            print(p["title"])
-    asyncio.run(_demo())
+    return out