Update mcp/cbio.py

mcp/cbio.py

@@ -1,24 +1,22 @@
-"""
-cbio.py – cBioPortal REST helper (no API-key required for public portal).
-Returns variant summary rows for a gene / profile.
-"""
-from functools import lru_cache
-import httpx, asyncio
+# mcp/cbio.py
 
-_BASE = "https://www.cbioportal.org/api/molecular-profiles/{profile}/genes/{gene}/mutations"
-_HEADERS = {"Accept": "application/json"}
+from pybioportal import PyBioPortal
+import asyncio
 
-@lru_cache(maxsize=256)
-async def fetch_cbio(
-    gene: str,
-    profile: str = "brca_tcga_pan_can_atlas_2018_mutations",
-) -> list[dict]:
-    if not gene:
-        return []
-    url = _BASE.format(profile=profile, gene=gene)
-    params = {"projection": "SUMMARY"}
-    async with httpx.AsyncClient(timeout=10, headers=_HEADERS) as c:
-        r = await c.get(url, params=params)
-        if r.status_code >= 400:
-            return []          # swallow 4xx/5xx → empty list
-        return r.json()
+async def fetch_cbio_variants(gene_symbol: str) -> list[dict]:
+    """
+    Fetches mutation/variant data for a specific gene symbol
+    from the default cBioPortal study (e.g. 'brca_mskcc_2020').
+    """
+    base_url = "https://www.cbioportal.org"
+    client = PyBioPortal(base_url)
+
+    # Build the query path for mutations in the gene
+    query_path = f"/mutation?gene={gene_symbol}&studyId=brca_mskcc_2020"
+    loop = asyncio.get_event_loop()
+    df = await loop.run_in_executor(
+        None,
+        lambda: client.get_data_frame(query_path)
+    )
+    # Convert DataFrame rows to dicts for JSON serialization
+    return df.to_dict(orient="records")