Update mcp/ctgov.py

mcp/ctgov.py

@@ -1,19 +1,19 @@
 #!/usr/bin/env python3
 """
-ClinicalTrials.gov helper – v2 primary, v1 fallback.
+ClinicalTrials.gov helper – v2 first, v1 fallback, 403-proof.
 """
 
 from __future__ import annotations
-import asyncio, httpx, urllib.parse
+import asyncio, httpx
 from functools import lru_cache
 from typing import List, Dict, Any
 
-# v2 REST
+# Endpoints
 _V2 = "https://clinicaltrials.gov/api/v2/studies"
-# legacy v1 (read-only but still works)
 _V1 = "https://clinicaltrials.gov/api/query/study_fields"
 
 _HEADERS = {
+    # Chrome-ish UA + explicit JSON accept header ← avoids 403
     "User-Agent": (
         "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
         "AppleWebKit/537.36 (KHTML, like Gecko) "
@@ -22,12 +22,14 @@ _HEADERS = {
     "Accept": "application/json",
 }
 _TIMEOUT = 15
-_MAX = 50  # hard cap
+_MAX     = 50          # cap page size
+
 
 async def _get(url: str, params: Dict[str, Any]) -> httpx.Response:
     async with httpx.AsyncClient(timeout=_TIMEOUT, headers=_HEADERS) as cli:
         return await cli.get(url, params=params)
 
+
 async def _try_v2(term: str, limit: int) -> List[Dict]:
     params = {
         "query"   : term,
@@ -37,9 +39,9 @@ async def _try_v2(term: str, limit: int) -> List[Dict]:
     r = await _get(_V2, params)
     if r.status_code == 200:
         return r.json().get("studies", [])
-    if r.status_code in {403, 429, 500, 503}:
-        raise RuntimeError("v2 unavailable")
-    r.raise_for_status()       # other 4xx
+    # let caller decide to retry or fallback
+    raise httpx.HTTPStatusError("v2 failed", request=r.request, response=r)
+
 
 async def _try_v1(term: str, limit: int) -> List[Dict]:
     params = {
@@ -50,22 +52,35 @@ async def _try_v1(term: str, limit: int) -> List[Dict]:
         "fmt"     : "json",
     }
     r = await _get(_V1, params)
-    r.raise_for_status()
-    return r.json().get("StudyFieldsResponse", {}).get("StudyFields", [])
+    if r.status_code == 200:
+        return r.json().get("StudyFieldsResponse", {}).get("StudyFields", [])
+    raise httpx.HTTPStatusError("v1 failed", request=r.request, response=r)
+
 
 @lru_cache(maxsize=512)
 async def search_trials(term: str, *, max_studies: int = 10) -> List[Dict]:
-    """
-    Return up to *max_studies* trial dicts; v2 first, v1 fallback.
-    """
+    """Return ≤ *max_studies* trials; silent `[]` if CT.gov is blocking."""
     limit = max(1, min(max_studies, _MAX))
-    # v2 with back-off: 2s → 4s
-    for delay in (0, 2, 4):
+
+    # try v2 with back-off
+    delay = 0
+    for _ in range(3):        # first + 2 retries
         try:
             if delay:
                 await asyncio.sleep(delay)
             return await _try_v2(term, limit)
-        except RuntimeError:
-            continue
-    # fallback to legacy endpoint
-    return await _try_v1(term, limit)
+        except httpx.HTTPStatusError as e:
+            if e.response.status_code not in {403, 429, 500, 503}:
+                break         # non-retryable
+            delay = 2 if delay == 0 else delay * 2
+
+    # fallback to v1 (once)
+    try:
+        return await _try_v1(term, limit)
+    except httpx.HTTPStatusError:
+        return []             # final graceful fallback
+
+
+# back-compat alias
+async def search_trials_v2(term: str, *, max_studies: int = 10):
+    return await search_trials(term, max_studies=max_studies)