Documentation-only follow-ups from the branch review

Author: sahilds1

server/api/views/assistant/eval_assistant.py

@@ -9,7 +9,8 @@
 # ///
 
 # uv script (or plain Python) to generate results to CSV, run from the terminal
-# Run from inside the container: docker compose exec backend python eval_assistant.py
+# Run from inside the container (working dir is /usr/src/server):
+#   docker compose exec backend python api/views/assistant/eval_assistant.py
 # 
 
 
@@ -20,6 +21,12 @@
 from concurrent.futures import ThreadPoolExecutor, as_completed
 
 # Django setup must come before any imports that touch the ORM
+# NOTE: from api/views/assistant/, "../../../../" resolves four levels up to
+# /usr/src (not /usr/src/server, where balancer_backend lives). So this insert
+# alone does not put the settings package on sys.path — running the script
+# relies on the container already having /usr/src/server on PYTHONPATH. Sanity-
+# check this the first time the eval is run for real; the path depth may need
+# adjusting (e.g. "../../../").
 sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "../../../../")))
 os.environ.setdefault("DJANGO_SETTINGS_MODULE", "balancer_backend.settings")
 

server/api/views/assistant/views.py

@@ -37,6 +37,12 @@ def post(self, request):
         try:
             user = request.user
 
+            # TODO: validate message and return a 400 when it is omitted or blank.
+            # @extend_schema documents message as required, but that schema is not
+            # enforced at runtime, so a missing/empty message reaches run_assistant
+            # and becomes the literal string "None" (str(None)) in the model input —
+            # producing confusing model behavior. Add a 400 to the responses schema
+            # when implementing.
             message = request.data.get("message", None)
             previous_response_id = request.data.get("previous_response_id", None)