Update genesis/ontology.py

genesis/ontology.py

@@ -1,39 +1,59 @@
 # genesis/ontology.py
 """
-Ontology Expansion for GENESIS-AI
-Uses UMLS & BioPortal to find related terms, synonyms, and hierarchies.
+Ontology Expansion & Merging for GENESIS-AI
+Uses UMLS & BioPortal to find related terms, synonyms, and hierarchies,
+and provides merging utilities for deduplication.
 """
 
 import os
 import requests
+import logging
+from typing import List, Dict, Union
+
+logging.basicConfig(level=logging.INFO)
 
 UMLS_API_KEY = os.getenv("UMLS_API_KEY")
 BIOPORTAL_API_KEY = os.getenv("BIOPORTAL_API_KEY")
 
-def expand_with_bioportal(term):
+
+# -------------------------
+# BioPortal Expansion
+# -------------------------
+def expand_with_bioportal(term: str) -> Union[List[str], Dict]:
     """Expand term using BioPortal ontology search."""
     try:
         url = f"https://data.bioontology.org/search?q={term}&apikey={BIOPORTAL_API_KEY}"
-        r = requests.get(url)
+        r = requests.get(url, timeout=20)
         r.raise_for_status()
         data = r.json()
         return list({res["prefLabel"] for res in data.get("collection", []) if "prefLabel" in res})
     except Exception as e:
+        logging.error(f"[Ontology] BioPortal expansion failed: {e}")
         return {"error": str(e)}
 
-def expand_with_umls(term):
+
+# -------------------------
+# UMLS Expansion
+# -------------------------
+def expand_with_umls(term: str) -> Union[List[str], Dict]:
     """Expand term using UMLS API."""
     try:
         url = f"https://uts-ws.nlm.nih.gov/rest/search/current?string={term}&apiKey={UMLS_API_KEY}"
-        r = requests.get(url)
+        r = requests.get(url, timeout=20)
         r.raise_for_status()
         data = r.json()
         return [res["name"] for res in data.get("result", {}).get("results", []) if "name" in res]
     except Exception as e:
+        logging.error(f"[Ontology] UMLS expansion failed: {e}")
         return {"error": str(e)}
 
-def expand_terms_with_ontology(term):
+
+# -------------------------
+# Combined Expansion
+# -------------------------
+def expand_terms_with_ontology(term: str) -> List[str]:
     """Combine BioPortal and UMLS expansions."""
+    logging.info(f"[Ontology] Expanding term: {term}")
     bioportal_terms = expand_with_bioportal(term)
     umls_terms = expand_with_umls(term)
     results = set()
@@ -41,4 +61,48 @@ def expand_terms_with_ontology(term):
         results.update(bioportal_terms)
     if isinstance(umls_terms, list):
         results.update(umls_terms)
-    return list(results) if results else [term]
+    final_terms = list(results) if results else [term]
+    logging.info(f"[Ontology] Found {len(final_terms)} expanded terms.")
+    return final_terms
+
+
+# -------------------------
+# Merge Ontology Terms
+# -------------------------
+def merge_ontology_terms(ontology_lists: List[List[Dict]]) -> List[Dict]:
+    """
+    Merge ontology term lists from multiple sources and remove duplicates.
+
+    Args:
+        ontology_lists: A list where each element is a list of ontology term dicts.
+            Example term: {"id": "GO:0008150", "label": "biological_process"}
+
+    Returns:
+        Merged list of unique ontology terms.
+    """
+    logging.info("[Ontology] Merging ontology term lists...")
+    merged_terms = {}
+
+    for source_list in ontology_lists:
+        for term in source_list:
+            if not isinstance(term, dict):
+                continue
+            term_id = term.get("id") or term.get("iri") or term.get("label")
+            if term_id and term_id not in merged_terms:
+                merged_terms[term_id] = {
+                    "id": term_id,
+                    "label": term.get("label", ""),
+                    "source": term.get("source", "unknown")
+                }
+
+    merged_list = list(merged_terms.values())
+    logging.info(f"[Ontology] Merged total {len(merged_list)} unique terms.")
+    return merged_list
+
+
+__all__ = [
+    "expand_with_bioportal",
+    "expand_with_umls",
+    "expand_terms_with_ontology",
+    "merge_ontology_terms"
+]