reformat

Author: penguine-ip

deepeval/evaluate/execute.py

@@ -1219,7 +1219,11 @@ async def _a_execute_agentic_test_case(
 
     test_case = LLMTestCase(
         input=golden.input,
-        actual_output=str(current_trace.output) if current_trace.output is not None else None,
+        actual_output=(
+            str(current_trace.output)
+            if current_trace.output is not None
+            else None
+        ),
         expected_output=current_trace.expected_output,
         context=current_trace.context,
         retrieval_context=current_trace.retrieval_context,

deepeval/openai_agents/callback_handler.py

@@ -53,7 +53,7 @@ def on_trace_start(self, trace: "Trace") -> None:
             _trace.metadata = make_json_serializable(_trace_metadata)
             current_trace_context.set(_trace)
 
-            trace_manager.add_span( # adds a dummy root span
+            trace_manager.add_span(  # adds a dummy root span
                 BaseSpan(
                     uuid=_trace_uuid,
                     trace_uuid=_trace_uuid,
@@ -77,7 +77,9 @@ def on_trace_end(self, trace: "Trace") -> None:
         _trace_metadata = trace_dict.get("metadata")
 
         if _thread_id or _trace_metadata:
-            trace_manager.remove_span(_trace_uuid) # removing the dummy root span
+            trace_manager.remove_span(
+                _trace_uuid
+            )  # removing the dummy root span
             trace_manager.end_trace(_trace_uuid)
             current_trace_context.set(None)
 

deepeval/openai_agents/runner.py

@@ -109,7 +109,7 @@ async def run(
             metric_collection=metric_collection,
             metrics=metrics,
             func_name="run",
-            function_kwargs={"input": input}, # also set below
+            function_kwargs={"input": input},  # also set below
         ) as observer:
             update_trace_attributes(
                 name=name,
@@ -177,7 +177,7 @@ def run_sync(
             metric_collection=metric_collection,
             metrics=metrics,
             func_name="run_sync",
-            function_kwargs={"input": input}, # also set below
+            function_kwargs={"input": input},  # also set below
         ) as observer:
             update_trace_attributes(
                 name=name,
@@ -263,7 +263,7 @@ def run_streamed(
         current_trace = current_trace_context.get()
         if not current_trace.input:
             current_trace.input = input
-            
+
         current_span = current_span_context.get()
         if current_span:
             current_span.input = input

tests/test_integrations/test_openai_agents/agents_app.py

@@ -107,42 +107,92 @@ async def run_weather_agent(user_input: str):
     )
     return result.final_output
 
+
 from agents import trace
 from multi_agents import triage_agent
+
+
 # with trace (group_id and metadata)
 async def main1():
-    with trace(workflow_name="test_workflow_1", group_id="test_group_id_1", metadata={"test_metadata_1": "test_metadata_1"}):
+    with trace(
+        workflow_name="test_workflow_1",
+        group_id="test_group_id_1",
+        metadata={"test_metadata_1": "test_metadata_1"},
+    ):
         user_query = "What's the weather like in London today?"
-        response_1 = await Runner.run(triage_agent, "Hola, ¿cómo estás?", metric_collection="test_collection_1", thread_id="test")
-        response_2 = await Runner.run(weather_agent, user_query, metric_collection="test_collection_1")
+        response_1 = await Runner.run(
+            triage_agent,
+            "Hola, ¿cómo estás?",
+            metric_collection="test_collection_1",
+            thread_id="test",
+        )
+        response_2 = await Runner.run(
+            weather_agent, user_query, metric_collection="test_collection_1"
+        )
         update_current_trace(input="initial input", output="final output")
 
+
 # without trace (group_id and metadata not present)
 async def main2():
     user_query = "What's the weather like in London today?"
-    response_1 = await Runner.run(triage_agent, "Hola, ¿cómo estás?", metric_collection="test_collection_1", thread_id="test")
-    response_2 = await Runner.run(weather_agent, user_query, metric_collection="test_collection_1")
-    
+    response_1 = await Runner.run(
+        triage_agent,
+        "Hola, ¿cómo estás?",
+        metric_collection="test_collection_1",
+        thread_id="test",
+    )
+    response_2 = await Runner.run(
+        weather_agent, user_query, metric_collection="test_collection_1"
+    )
+
 
 async def main3():
     user_query = "What's the weather like in London today?"
-    with trace(workflow_name="test_workflow_1", group_id="test_group_id_1", metadata={"test_metadata_1": "test_metadata_1"}):
-        response_2 = await Runner.run(weather_agent, user_query, metric_collection="test_collection_1")
-    with trace(workflow_name="test_workflow_2", group_id="test_group_id_2", metadata={"test_metadata_2": "test_metadata_2"}):
-        response_1 = await Runner.run(triage_agent, "Hola, ¿cómo estás?", metric_collection="test_collection_1", thread_id="test")
+    with trace(
+        workflow_name="test_workflow_1",
+        group_id="test_group_id_1",
+        metadata={"test_metadata_1": "test_metadata_1"},
+    ):
+        response_2 = await Runner.run(
+            weather_agent, user_query, metric_collection="test_collection_1"
+        )
+    with trace(
+        workflow_name="test_workflow_2",
+        group_id="test_group_id_2",
+        metadata={"test_metadata_2": "test_metadata_2"},
+    ):
+        response_1 = await Runner.run(
+            triage_agent,
+            "Hola, ¿cómo estás?",
+            metric_collection="test_collection_1",
+            thread_id="test",
+        )
+
 
 async def main4():
     user_query = "What's the weather like in London today?"
-    with trace(workflow_name="test_workflow_1", group_id="test_group_id_1", metadata={"test_metadata_1": "test_metadata_1"}):
-        run_streamed_1 = Runner.run_streamed(weather_agent, user_query, metric_collection="test_collection_1")
+    with trace(
+        workflow_name="test_workflow_1",
+        group_id="test_group_id_1",
+        metadata={"test_metadata_1": "test_metadata_1"},
+    ):
+        run_streamed_1 = Runner.run_streamed(
+            weather_agent, user_query, metric_collection="test_collection_1"
+        )
         async for chunk in run_streamed_1.stream_events():
             print(chunk, end="", flush=True)
             print("=" * 50)
-        run_streamed_2 = Runner.run_streamed(triage_agent, "Hola, ¿cómo estás?", metric_collection="test_collection_1", thread_id="test")
+        run_streamed_2 = Runner.run_streamed(
+            triage_agent,
+            "Hola, ¿cómo estás?",
+            metric_collection="test_collection_1",
+            thread_id="test",
+        )
         async for chunk in run_streamed_2.stream_events():
             print(chunk, end="", flush=True)
             print("=" * 50)
 
+
 def execute_agent():
     asyncio.run(main1())
     # asyncio.run(main2())

tests/test_integrations/test_openai_agents/multi_agents.py

@@ -1,6 +1,7 @@
 import asyncio
 from deepeval.openai_agents import Agent, Runner
 from deepeval.prompt import Prompt
+
 # from deepeval.openai_agents import DeepEvalTracingProcessor
 
 # from agents import add_trace_processor