silence warnings for demo

Author: Andrei Bratu

src/humanloop/eval_utils/run.py

@@ -23,6 +23,7 @@
 from functools import partial
 from logging import INFO
 from typing import Callable, Dict, List, Literal, Optional, Sequence, Tuple, TypeVar, Union
+import warnings
 
 from humanloop import EvaluatorResponse, FlowResponse, PromptResponse, ToolResponse
 from humanloop.core.api_error import ApiError
@@ -121,13 +122,23 @@ def _overload_call(self, **kwargs) -> PromptCallResponse:
                 # TODO: Bug found in backend: not specifying a model 400s but creates a File
                 raise HumanloopUtilityError(message=str(e)) from e
 
+            response_copy = response.dict()
             prompt_utility_context = get_prompt_utility_context()
+            for idx, _ in enumerate(response_copy.get("logs", [])):
+                del response_copy["logs"][idx]["created_at"]
+            for idx, _ in enumerate(response_copy["prompt"].get("environment", [])):
+                del response_copy["prompt"]["environment"][idx]["created_at"]
+            del response_copy["prompt"]["last_used_at"]
+            del response_copy["prompt"]["updated_at"]
+            del response_copy["prompt"]["created_at"]
+            del response_copy["start_time"]
+            del response_copy["end_time"]
 
             with prompt_utility_context.tracer.start_as_current_span(HUMANLOOP_INTERCEPTED_HL_CALL_SPAN_NAME) as span:
                 write_to_opentelemetry_span(
                     span=span,
                     key=HUMANLOOP_INTERCEPTED_HL_CALL_RESPONSE,
-                    value=response.dict(),
+                    value=response_copy,
                 )
             return response
         else:

src/humanloop/otel/exporter.py

@@ -375,6 +375,7 @@ def _mark_trace_complete_if_needed(self, trace_head_span_id: int):
 
     def _keep_track_of_trace(self, log_id: str, parent_log_id: str):
         for trace in self._traces:
+            found = False
             if parent_log_id in trace:
                 trace.add(log_id)
                 found = True

tests/integration/evaluate_medqa/conftest.py

@@ -173,7 +173,7 @@ def ask_question(**inputs) -> str:
             return call_model(**inputs)
 
         with open(medqa_dataset_path, "r") as file:
-            datapoints = [json.loads(line) for line in file.readlines()][:20]
+            datapoints = [json.loads(line) for line in file.readlines()]
 
         for path, code, return_type in [
             (get_test_path("Levenshtein Distance"), levenshtein, "number"),

tests/integration/evaluate_medqa/test_evaluate_medqa.py

@@ -8,8 +8,8 @@
 from humanloop import Humanloop
 
 
-@pytest.mark.skip("skip for demo")
-@pytest.mark.parametrize("use_call", [False])
+# @pytest.mark.skip("skip for demo")
+@pytest.mark.parametrize("use_call", [True])
 def test_scenario(
     evaluate_medqa_scenario_factory: Callable[[bool], MedQAScenario],
     humanloop_client: Humanloop,
@@ -22,8 +22,6 @@ def test_scenario(
     levenshtein_path = evaluate_medqa_scenario.levenshtein_path
     exact_match_path = evaluate_medqa_scenario.exact_match_path
 
-    assert len(medqa_dataset) == 20
-
     humanloop_client.evaluations.run(  # type: ignore [attr-defined]
         name="Test",
         file={