Update genesis/api_clients/pubmed_api.py

genesis/api_clients/pubmed_api.py

@@ -1,34 +1,52 @@
 # genesis/api_clients/pubmed_api.py
 import requests
-import os
-
-NCBI_API_KEY = os.getenv("NCBI_API_KEY")  # Stored in HF Secrets
-BASE_URL = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"
+from xml.etree import ElementTree as ET
 
 def search_pubmed(query: str, max_results: int = 10):
-    """Search PubMed articles."""
-    url = f"{BASE_URL}/esearch.fcgi"
-    params = {
+    """
+    Search PubMed using the NCBI E-utilities API.
+    Returns a list of dicts with 'title', 'authors', 'pub_date', 'link'.
+    """
+    base_url = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi"
+    summary_url = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esummary.fcgi"
+
+    # Step 1: Search PubMed IDs
+    search_params = {
         "db": "pubmed",
         "term": query,
         "retmax": max_results,
-        "retmode": "json",
-        "api_key": NCBI_API_KEY
+        "retmode": "json"
     }
-    r = requests.get(url, params=params)
-    r.raise_for_status()
-    return r.json()
+    search_res = requests.get(base_url, params=search_params)
+    search_res.raise_for_status()
+    id_list = search_res.json().get("esearchresult", {}).get("idlist", [])
+
+    if not id_list:
+        return []
 
-def fetch_pubmed_details(id_list):
-    """Fetch article details for given PubMed IDs."""
-    ids = ",".join(id_list)
-    url = f"{BASE_URL}/efetch.fcgi"
-    params = {
+    # Step 2: Fetch summaries for IDs
+    summary_params = {
         "db": "pubmed",
-        "id": ids,
-        "retmode": "xml",
-        "api_key": NCBI_API_KEY
+        "id": ",".join(id_list),
+        "retmode": "xml"
     }
-    r = requests.get(url, params=params)
-    r.raise_for_status()
-    return r.text
+    summary_res = requests.get(summary_url, params=summary_params)
+    summary_res.raise_for_status()
+
+    root = ET.fromstring(summary_res.text)
+    papers = []
+    for docsum in root.findall(".//DocSum"):
+        paper = {"title": None, "authors": [], "pub_date": None, "link": None}
+        for item in docsum.findall("Item"):
+            if item.attrib.get("Name") == "Title":
+                paper["title"] = item.text
+            elif item.attrib.get("Name") == "PubDate":
+                paper["pub_date"] = item.text
+            elif item.attrib.get("Name") == "AuthorList":
+                paper["authors"] = [author.text for author in item.findall("Item")]
+        uid_elem = docsum.find("Id")
+        if uid_elem is not None:
+            paper["link"] = f"https://pubmed.ncbi.nlm.nih.gov/{uid_elem.text}/"
+        papers.append(paper)
+
+    return papers