tests: improve the time for some tests (#2592)

Author: Kludex

tests/providers/cassettes/test_openrouter/test_openrouter_with_google_model.yaml

@@ -1,100 +1,4 @@
 interactions:
-- request:
-    headers:
-      accept:
-      - application/json
-      accept-encoding:
-      - gzip, deflate
-      connection:
-      - keep-alive
-      content-length:
-      - '164'
-      content-type:
-      - application/json
-      host:
-      - openrouter.ai
-    method: POST
-    parsed_body:
-      messages:
-      - content: Be helpful.
-        role: system
-      - content: Tell me a joke.
-        role: user
-      model: google/gemini-2.0-flash-exp:free
-      n: 1
-      stream: false
-    uri: https://openrouter.ai/api/v1/chat/completions
-  response:
-    headers:
-      access-control-allow-origin:
-      - '*'
-      connection:
-      - keep-alive
-      content-length:
-      - '242'
-      content-type:
-      - application/json
-      vary:
-      - Accept-Encoding
-    parsed_body:
-      error:
-        code: 429
-        message: Provider returned error
-        metadata:
-          provider_name: Google
-          raw: google/gemini-2.0-flash-exp:free is temporarily rate-limited upstream; please retry shortly.
-      user_id: user_2uRh0l3Yi3hdjBArTOSmLXWJBc4
-    status:
-      code: 429
-      message: Too Many Requests
-- request:
-    headers:
-      accept:
-      - application/json
-      accept-encoding:
-      - gzip, deflate
-      connection:
-      - keep-alive
-      content-length:
-      - '164'
-      content-type:
-      - application/json
-      host:
-      - openrouter.ai
-    method: POST
-    parsed_body:
-      messages:
-      - content: Be helpful.
-        role: system
-      - content: Tell me a joke.
-        role: user
-      model: google/gemini-2.0-flash-exp:free
-      n: 1
-      stream: false
-    uri: https://openrouter.ai/api/v1/chat/completions
-  response:
-    headers:
-      access-control-allow-origin:
-      - '*'
-      connection:
-      - keep-alive
-      content-length:
-      - '252'
-      content-type:
-      - application/json
-      vary:
-      - Accept-Encoding
-    parsed_body:
-      error:
-        code: 429
-        message: Provider returned error
-        metadata:
-          provider_name: Google AI Studio
-          raw: google/gemini-2.0-flash-exp:free is temporarily rate-limited upstream; please retry shortly.
-      user_id: user_2uRh0l3Yi3hdjBArTOSmLXWJBc4
-    status:
-      code: 429
-      message: Too Many Requests
 - request:
     headers:
       accept:

tests/test_agent.py

@@ -4074,9 +4074,7 @@ async def test_implicit_context_manager():
     toolset = CombinedToolset([server1, PrefixedToolset(server2, 'prefix')])
     agent = Agent('test', toolsets=[toolset])
 
-    async with agent.iter(
-        user_prompt='Hello',
-    ):
+    async with agent.iter(user_prompt='Hello'):
         assert server1.is_running
         assert server2.is_running
 

tests/test_mcp.py

@@ -178,7 +178,6 @@ def test_sse_server_conflicting_timeout_params():
         )
 
 
-@pytest.mark.vcr()
 async def test_agent_with_stdio_server(allow_model_requests: None, agent: Agent):
     async with agent:
         result = await agent.run('What is 0 degrees Celsius in Fahrenheit?')
@@ -280,7 +279,6 @@ def get_none() -> None:  # pragma: no cover
             await agent.run('No conflict')
 
 
-@pytest.mark.vcr()
 async def test_agent_with_server_not_running(agent: Agent, allow_model_requests: None):
     result = await agent.run('What is 0 degrees Celsius in Fahrenheit?')
     assert result.output == snapshot('0 degrees Celsius is 32.0 degrees Fahrenheit.')
@@ -306,7 +304,6 @@ async def test_log_level_set(run_context: RunContext[int]):
         assert result == snapshot('info')
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_str(allow_model_requests: None, agent: Agent):
     async with agent:
         result = await agent.run('What is the weather in Mexico City?')
@@ -379,7 +376,6 @@ async def test_tool_returning_str(allow_model_requests: None, agent: Agent):
         )
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_text_resource(allow_model_requests: None, agent: Agent):
     async with agent:
         result = await agent.run('Get me the product name')
@@ -446,7 +442,6 @@ async def test_tool_returning_text_resource(allow_model_requests: None, agent: A
         )
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_text_resource_link(allow_model_requests: None, agent: Agent):
     async with agent:
         result = await agent.run('Get me the product name via get_product_name_link')
@@ -513,7 +508,6 @@ async def test_tool_returning_text_resource_link(allow_model_requests: None, age
         )
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_image_resource(allow_model_requests: None, agent: Agent, image_content: BinaryContent):
     async with agent:
         result = await agent.run('Get me the image resource')
@@ -587,7 +581,6 @@ async def test_tool_returning_image_resource(allow_model_requests: None, agent:
         )
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_image_resource_link(
     allow_model_requests: None, agent: Agent, image_content: BinaryContent
 ):
@@ -663,7 +656,6 @@ async def test_tool_returning_image_resource_link(
         )
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_audio_resource(
     allow_model_requests: None, agent: Agent, audio_content: BinaryContent, gemini_api_key: str
 ):
@@ -712,7 +704,6 @@ async def test_tool_returning_audio_resource(
         )
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_audio_resource_link(
     allow_model_requests: None, agent: Agent, audio_content: BinaryContent, gemini_api_key: str
 ):
@@ -771,7 +762,6 @@ async def test_tool_returning_audio_resource_link(
         )
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_image(allow_model_requests: None, agent: Agent, image_content: BinaryContent):
     async with agent:
         result = await agent.run('Get me an image')
@@ -845,7 +835,6 @@ async def test_tool_returning_image(allow_model_requests: None, agent: Agent, im
         )
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_dict(allow_model_requests: None, agent: Agent):
     async with agent:
         result = await agent.run('Get me a dict, respond on one line')
@@ -906,7 +895,6 @@ async def test_tool_returning_dict(allow_model_requests: None, agent: Agent):
         )
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_error(allow_model_requests: None, agent: Agent):
     async with agent:
         result = await agent.run('Get me an error, pass False as a value, unless the tool tells you otherwise')
@@ -1011,7 +999,6 @@ async def test_tool_returning_error(allow_model_requests: None, agent: Agent):
         )
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_none(allow_model_requests: None, agent: Agent):
     async with agent:
         result = await agent.run('Call the none tool and say Hello')
@@ -1072,7 +1059,6 @@ async def test_tool_returning_none(allow_model_requests: None, agent: Agent):
         )
 
 
-@pytest.mark.vcr()
 async def test_tool_returning_multiple_items(allow_model_requests: None, agent: Agent, image_content: BinaryContent):
     async with agent:
         result = await agent.run('Get me multiple items and summarize in one sentence')

tests/test_tenacity.py

@@ -553,7 +553,7 @@ def validate_response(response: httpx.Response):
 
         controller = Retrying(
             retry=retry_if_exception_type(httpx.HTTPStatusError),
-            wait=wait_retry_after(max_wait=2),  # Cap at 2 seconds for tests
+            wait=wait_retry_after(max_wait=0.1),  # Cap at 0.1 seconds for tests
             stop=stop_after_attempt(3),
             reraise=True,
         )
@@ -568,5 +568,6 @@ def validate_response(response: httpx.Response):
 
         assert result is mock_response_success
         assert mock_transport.handle_request.call_count == 2
-        # Should have waited approximately 2 seconds (capped by max_wait)
-        assert 1.8 <= (end_time - start_time) <= 3.0
+        # Should have waited approximately 0.2 seconds (capped by max_wait)
+        duration = end_time - start_time
+        assert 0.1 <= duration <= 0.2