Update genesis/api_clients/bioportal_api.py

genesis/api_clients/bioportal_api.py

@@ -1,78 +1,69 @@
 # genesis/api_clients/bioportal_api.py
 import requests
-import os
-from urllib.parse import quote
+from typing import List, Dict, Optional
 
-BIOPORTAL_API_KEY = os.getenv("BIOPORTAL_API_KEY")  # Store this in Hugging Face secrets
 BIOPORTAL_BASE_URL = "https://data.bioontology.org"
+BIOPORTAL_API_KEY = None  # Load from secrets in Hugging Face or local .env
 
-if not BIOPORTAL_API_KEY:
-    raise ValueError("BIOPORTAL_API_KEY not found in environment variables.")
-
-HEADERS = {
-    "Authorization": f"apikey token={BIOPORTAL_API_KEY}"
-}
-
-
-def search_concept(term: str, ontology: str = None, max_results: int = 10):
+def search_term(term: str, ontologies: Optional[List[str]] = None, max_results: int = 20) -> List[Dict]:
     """
-    Search for a biomedical concept across BioPortal ontologies.
-    
-    Args:
-        term (str): The search term (e.g., "BRCA1", "glioblastoma").
-        ontology (str): Optional ontology acronym (e.g., "NCIT", "SNOMEDCT").
-        max_results (int): Maximum number of results.
+    Search BioPortal for a term across one or more ontologies.
+    If no ontologies specified, searches all available ones.
     """
+    headers = {"Authorization": f"apikey token={BIOPORTAL_API_KEY}"}
     params = {
         "q": term,
         "pagesize": max_results
     }
-    if ontology:
-        params["ontology"] = ontology
-
-    url = f"{BIOPORTAL_BASE_URL}/search"
-    response = requests.get(url, headers=HEADERS, params=params)
-    response.raise_for_status()
-    return response.json().get("collection", [])
+    if ontologies:
+        params["ontologies"] = ",".join(ontologies)
 
+    r = requests.get(f"{BIOPORTAL_BASE_URL}/search", headers=headers, params=params)
+    r.raise_for_status()
+    data = r.json()
 
-def get_concept_details(ontology_acronym: str, concept_id: str):
-    """
-    Fetch detailed information about a concept given ontology acronym and concept ID.
-    """
-    encoded_id = quote(concept_id, safe="")
-    url = f"{BIOPORTAL_BASE_URL}/ontologies/{ontology_acronym}/classes/{encoded_id}"
-    response = requests.get(url, headers=HEADERS)
-    response.raise_for_status()
-    return response.json()
+    results = []
+    for result in data.get("collection", []):
+        results.append({
+            "pref_label": result.get("prefLabel"),
+            "definition": result.get("definition", ["No definition found"])[0],
+            "synonyms": result.get("synonym", []),
+            "ontology": result.get("links", {}).get("ontology"),
+            "iri": result.get("@id"),
+            "score": result.get("score", 0)
+        })
+    return results
 
 
-def get_ontologies():
+def get_concept_details(ontology_acronym: str, concept_id: str) -> Dict:
     """
-    List all available ontologies in BioPortal.
+    Fetch detailed metadata for a specific ontology concept.
     """
-    url = f"{BIOPORTAL_BASE_URL}/ontologies"
-    response = requests.get(url, headers=HEADERS)
-    response.raise_for_status()
-    return [
-        {"acronym": o.get("acronym"), "name": o.get("name"), "description": o.get("description")}
-        for o in response.json()
-    ]
+    headers = {"Authorization": f"apikey token={BIOPORTAL_API_KEY}"}
+    r = requests.get(f"{BIOPORTAL_BASE_URL}/ontologies/{ontology_acronym}/classes/{concept_id}",
+                     headers=headers)
+    r.raise_for_status()
+    data = r.json()
+
+    return {
+        "pref_label": data.get("prefLabel"),
+        "definition": data.get("definition", ["No definition found"])[0],
+        "synonyms": data.get("synonym", []),
+        "properties": data.get("properties", {}),
+        "relations": data.get("links", {})
+    }
 
 
-def map_term_across_ontologies(term: str, ontologies: list = None):
+def map_to_ontologies(term: str, ontologies: List[str]) -> Dict[str, List[str]]:
     """
-    Map a term across multiple ontologies to unify synonyms and cross-references.
+    Map a given term to equivalent terms in multiple ontologies.
+    Returns a dictionary mapping ontology → list of equivalent labels.
     """
-    mappings = {}
-    target_ontologies = ontologies or [o["acronym"] for o in get_ontologies()]
-    
-    for ont in target_ontologies:
-        results = search_concept(term, ontology=ont, max_results=1)
-        if results:
-            mappings[ont] = {
-                "label": results[0].get("prefLabel"),
-                "uri": results[0].get("@id"),
-                "synonyms": results[0].get("synonym", []),
-            }
-    return mappings
+    results = search_term(term, ontologies=ontologies)
+    mapping = {}
+    for res in results:
+        onto_name = res.get("ontology", "Unknown")
+        if onto_name not in mapping:
+            mapping[onto_name] = []
+        mapping[onto_name].append(res.get("pref_label"))
+    return mapping