[FIX] fix long urls (#1085)

* make okay fix

* fix style

* update openapi

Author: jdkent

store/backend/neurostore/openapi

@@ -441,14 +441,27 @@ def clear_cache(unique_ids):
 
 def cache_key_creator(*args, **kwargs):
     # relevant pieces of information
-    # 1. the query arguments
+    # 1. the query arguments (including extra_args if present)
     # 2. the path
     # 3. the user
     path = request.path
     user = get_current_user().id if get_current_user() else ""
-    args_as_sorted_tuple = tuple(
-        sorted(pair for pair in request.args.items(multi=True))
-    )
+
+    # Get query args from request
+    query_items = list(request.args.items(multi=True))
+
+    # If extra_args is present, merge into query_items
+    extra_args = kwargs.get("extra_args")
+    if extra_args:
+        for k, v in extra_args.items():
+            # Support both single values and lists
+            if isinstance(v, list):
+                for item in v:
+                    query_items.append((k, item))
+            else:
+                query_items.append((k, v))
+
+    args_as_sorted_tuple = tuple(sorted(query_items))
     query_args = str(args_as_sorted_tuple)
 
     cache_key = "_".join([path, query_args, user])
@@ -628,9 +641,14 @@ def create_metadata(self, q, total):
         return {"total_count": total}
 
     @cache.cached(60 * 60, query_string=True, make_cache_key=cache_key_creator)
-    def search(self):
+    def search(self, extra_args=None):
         # Parse arguments using webargs
+        import logging
+
+        logging.warning(f"I RAN HERE Request args: {request.args}")
         args = parser.parse(self._user_args, request, location="query")
+        if extra_args:
+            args.update(extra_args)
 
         m = self._model  # for brevity
         q = m.query

store/backend/neurostore/resources/base.py

@@ -1,7 +1,7 @@
 """Pipeline related resources"""
 
 from sqlalchemy import text, and_, or_
-from flask import abort
+from flask import abort, request
 
 from sqlalchemy.orm import selectinload, aliased
 from webargs import fields
@@ -272,3 +272,22 @@ def eager_load(self, q, args=None):
             )
         )
         return q
+
+    def post(self):
+        """
+        If 'study_ids' is present in the request body, treat as a search (bypass authorization).
+        Only study_ids are in the body; all other filters are in the query string.
+        Otherwise, treat as a creation (require authorization).
+        """
+        data = request.get_json() or {}
+
+        if "study_ids" in data:
+            # Bypass authorization for search requests
+            # Convert study_ids to study_id for consistent filtering
+            study_ids = data.get("study_ids", [])
+            extra_args = {"study_id": study_ids}
+            # Call cached search (enables cache for study_ids POST)
+            return self.search(extra_args=extra_args)
+        else:
+            # Standard POST: require authorization (enforced by OpenAPI and Flask)
+            return super().post(self)

store/backend/neurostore/resources/pipeline.py

@@ -590,3 +590,26 @@ def test_pipeline_config_with_schema(auth_client, pipeline1):
     assert data["config_args"]["extractor"] == "ParticipantDemographicsExtractor"
     assert "extractor_kwargs" in data["config_args"]
     assert data["config_args"]["extraction_model"] == "gpt-4"
+
+
+def test_post_pipeline_study_results_with_study_ids(
+    auth_client, result1, result2, pipeline_study_result_payload
+):
+    study1_id = pipeline_study_result_payload[0]["base_study_id"]
+    study2_id = pipeline_study_result_payload[1]["base_study_id"]
+    url = "/api/pipeline-study-results/"
+    payload = {"study_ids": [study1_id, study2_id]}
+
+    # First request: should call search logic and return both results
+    response = auth_client.post(url, data=payload, content_type="application/json")
+    assert response.status_code == 200
+    data = response.json()
+    returned_ids = {r["base_study_id"] for r in data["results"]}
+    assert returned_ids == {study1_id, study2_id}
+
+    # Test cache hit (currently not working)
+    response2 = auth_client.post(url, data=payload)
+    assert response2.status_code == 200
+    data2 = response2.json()
+    returned_ids2 = {r["base_study_id"] for r in data2["results"]}
+    assert returned_ids2 == {study1_id, study2_id}