Apply pre-commit fixes

Author: ulgens

.coderabbit.yaml

@@ -4,9 +4,9 @@ reviews:
     - path: "/**/test_*.py"
       instructions: |
         Review the following unit test code written using the pytest test library. Ensure that:
-        - The code adheres to best practices associated with Pytest. The project uses uv.  
+        - The code adheres to best practices associated with Pytest. The project uses uv.
         - Descriptive test names are used to clearly convey the intent of each test.
-        - For Integration tests, we should avoid mocking. Integration tests of classes should use the actual client. 
-        - The test should convey actual value. Tests which initialise a class and test the values of those classes should be discouraged 
-        - Discourage verbosity and redundant tests    
-tone_instructions: "If you must write a poem, write it in the style of Sylvia Plath."
+        - For Integration tests, we should avoid mocking. Integration tests of classes should use the actual client.
+        - The test should convey actual value. Tests which initialise a class and test the values of those classes should be discouraged
+        - Discourage verbosity and redundant tests
+tone_instructions: "If you must write a poem, write it in the style of Sylvia Plath."

.cursor/rules/python.mdc

@@ -4,8 +4,8 @@ globs:
 alwaysApply: true
 ---
 
-never mock. 
-tests use pytest. 
+never mock.
+tests use pytest.
 never adjust sys.path.
 
 never write: except Exception as e

.github/workflows/release.yml

@@ -56,4 +56,4 @@ jobs:
 
       - name: Install core PyPI using uv
         run: |
-          uv pip install vision-agents==${{ needs.build-core.outputs.version }}
+          uv pip install vision-agents==${{ needs.build-core.outputs.version }}

PRODUCTION.md

@@ -17,4 +17,4 @@ A distributed approach where agents are available in many regions is ideal
 
 - Fly
 - Digital Ocean
-- AWS
+- AWS

README.md

@@ -225,7 +225,7 @@ While building the integrations, here are the limitations we've noticed (Dec 202
 * Longer videos can cause the AI to lose context. For instance if it's watching a soccer match it will get confused after 30 seconds
 * Most applications require a combination of small specialized models like Yolo/Roboflow/Moondream, API calls to get more context and larger models like gemini/openAI
 * Image size & FPS need to stay relatively low due to performance constraints
-* Video doesn’t trigger responses in realtime models. You always need to send audio/text to trigger a response. 
+* Video doesn’t trigger responses in realtime models. You always need to send audio/text to trigger a response.
 
 ## Star History
 

SECURITY.md

@@ -25,5 +25,5 @@ While we appreciate any information that you are willing to provide, please make
 
 ### Scope
 
-Only code in this repository is in scope.  
+Only code in this repository is in scope.
 Third-party services (hosted demo, npm registry, etc.) are handled separately.

agents-core/README.md

@@ -10,4 +10,4 @@ Build Vision Agents quickly with any model or video provider.
 
 Created by Stream, uses [Stream's edge network](https://getstream.io/video/) for ultra-low latency.
 
-See [Github](https://github.com/GetStream/Vision-Agents).
+See [Github](https://github.com/GetStream/Vision-Agents).

agents-core/vision_agents/PROTOBUF_GENERATION.md

@@ -59,7 +59,7 @@ class Participant(DataClassJsonMixin):
     is_speaking: Optional[bool] = None
     audio_level: Optional[float] = None
     # ... all other fields
-    
+
     @classmethod
     def from_proto(cls, proto_obj) -> 'Participant':
         """Create from protobuf Participant."""
@@ -243,7 +243,7 @@ The EventManager has been updated to seamlessly handle the new protobuf events:
    ```python
    from vision_agents.core.events.manager import EventManager
    from vision_agents.core.edge.sfu_events import AudioLevelEvent
-   
+
    manager = EventManager()
    manager.register(AudioLevelEvent)
    ```
@@ -255,13 +255,13 @@ The EventManager has been updated to seamlessly handle the new protobuf events:
      event = AudioLevelEvent.from_proto(proto, session_id='session123')
      manager.send(event)  # BaseEvent fields preserved
      ```
-   
+
    - Send raw protobuf messages (auto-wrapped):
      ```python
      proto = events_pb2.AudioLevel(user_id='user456', level=0.95)
      manager.send(proto)  # Automatically wrapped in AudioLevelEvent
      ```
-   
+
    - Create events without payload (all fields optional):
      ```python
      event = AudioLevelEvent()  # No protobuf payload needed
@@ -283,4 +283,3 @@ The EventManager has been updated to seamlessly handle the new protobuf events:
 - **Simplified logic**: Single check distinguishes raw protobuf from wrapped events
 - **Type safety**: All generated events properly inherit from BaseEvent
 - **Flexible usage**: Use raw protobuf or wrapped events interchangeably
-

docs/ai/instructions/ai-events-example.md

@@ -47,7 +47,7 @@ class MyPlugin:
     def __init__(self):
         # EventManager is automatically available
         self.events = EventManager()
-        
+
         # Register your custom events
         self.events.register_events_from_module(events)
 ```
@@ -121,7 +121,7 @@ class MyValidatedEvent(PluginBaseEvent):
     type: str = field(default='plugin.myplugin.validated', init=False)
     text: str = ""
     confidence: float = 0.0
-    
+
     def __post_init__(self):
         if not self.text:
             raise ValueError("Text cannot be empty")
@@ -160,17 +160,17 @@ class MyPlugin:
                 message="Processing started",
                 data=data
             ))
-            
+
             # Do processing
             result = await self._process(data)
-            
+
             # Emit success event
             self.events.send(MyPluginEvent(
                 plugin_name="myplugin",
                 message="Processing completed",
                 data=result
             ))
-            
+
         except Exception as e:
             # Emit error event
             self.events.send(MyPluginErrorEvent(
@@ -205,21 +205,21 @@ class MyPlugin:
         super().__init__()
         self.events.register_events_from_module(events)
         self._setup_event_handlers()
-    
+
     def _setup_event_handlers(self):
         """Set up event handlers for the plugin."""
-        
+
         @self.events.subscribe
         async def handle_stt_transcript(event: STTTranscriptEvent):
             """Handle speech-to-text transcripts."""
             if event.is_final:
                 await self._process_transcript(event.text)
-        
+
         @self.events.subscribe
         async def handle_llm_response(event: LLMResponseEvent):
             """Handle LLM responses."""
             await self._process_llm_response(event.text)
-        
+
         @self.events.subscribe
         async def handle_error_events(event: MyPluginErrorEvent | STTErrorEvent):
             """Handle error events."""
@@ -268,16 +268,16 @@ class MyPlugin(PluginBase):
         super().__init__()
         # Register custom events
         self.events.register_events_from_module(events)
-    
+
     async def process(self, data):
         # Send custom events
         self.events.send(events.MyPluginStartEvent(
             plugin_name="myplugin",
             config=self.config
         ))
-        
+
         result = await self._process_data(data)
-        
+
         self.events.send(events.MyPluginDataEvent(
             plugin_name="myplugin",
             data=result,
@@ -296,21 +296,21 @@ class SimpleSTT(STT):
     def __init__(self):
         super().__init__()
         # No need to register custom events - use base class events
-    
+
     async def transcribe(self, audio_data: bytes) -> str:
         try:
             result = await self._call_api(audio_data)
-            
+
             # Send base class event
             self.events.send(STTTranscriptEvent(
                 plugin_name="simple_stt",
                 text=result.text,
                 confidence=result.confidence,
                 is_final=True
             ))
-            
+
             return result.text
-            
+
         except Exception as e:
             # Send error event
             self.events.send(STTErrorEvent(
@@ -332,32 +332,32 @@ class MyAgent(Agent):
     def __init__(self, **kwargs):
         super().__init__(**kwargs)
         self._setup_agent_handlers()
-    
+
     def _setup_agent_handlers(self):
         @self.events.subscribe
         async def handle_agent_say(event: AgentSayEvent):
             """Handle when agent wants to say something."""
             print(f"Agent wants to say: {event.text}")
-            
+
             # Process the speech request
             await self._process_speech_request(event)
-    
+
     # Three ways to send events:
-    
+
     # Method 1: Direct event sending
     def send_custom_event(self, data):
         self.events.send(MyCustomEvent(
             plugin_name="agent",
             data=data
         ))
-    
+
     # Method 2: Convenience method
     def send_event_convenience(self, data):
         self.send(MyCustomEvent(
             plugin_name="agent",
             data=data
         ))
-    
+
     # Method 3: High-level speech
     async def make_agent_speak(self, text):
         await self.say(text, metadata={"source": "custom_handler"})
@@ -423,7 +423,7 @@ class ValidatedEvent(PluginBaseEvent):
     type: str = field(default='plugin.myplugin.validated', init=False)
     text: str = ""
     confidence: float = 0.0
-    
+
     def __post_init__(self):
         if not self.text.strip():
             raise ValueError("Text cannot be empty")
@@ -440,20 +440,20 @@ from my_plugin import MyPlugin
 @pytest.mark.asyncio
 async def test_plugin_events():
     plugin = MyPlugin()
-    
+
     # Track events
     events_received = []
-    
+
     @plugin.events.subscribe
     async def track_events(event):
         events_received.append(event)
-    
+
     # Trigger event
     await plugin.process_data("test")
-    
+
     # Wait for events
     await plugin.events.wait()
-    
+
     # Verify events
     assert len(events_received) > 0
     assert any(isinstance(e, MyPluginEvent) for e in events_received)
@@ -483,15 +483,15 @@ class OpenAILLM(LLM):
         super().__init__()
         self.events.register_events_from_module(events)
         self.model = model
-    
+
     def _standardize_and_emit_event(self, event: ResponseStreamEvent):
         # Send raw OpenAI event
         self.events.send(events.OpenAIStreamEvent(
             plugin_name="openai",
             event_type=event.type,
             event_data=event
         ))
-        
+
         if event.type == "response.error":
             self.events.send(events.LLMErrorEvent(
                 plugin_name="openai",
@@ -519,10 +519,10 @@ class Realtime(realtime.Realtime):
     def __init__(self, **kwargs):
         super().__init__(**kwargs)
         self.events.register_events_from_module(events)
-    
+
     async def connect(self):
         # ... connection logic ...
-        
+
         # Emit connection event
         self.events.send(events.GeminiConnectedEvent(
             plugin_name="gemini",

docs/ai/instructions/ai-llm.md

@@ -14,19 +14,19 @@ class MyLLM(LLM):
         super().__init__()
         self.model = model
         self.client = client
-        
-        
+
+
     # native method wrapped. wrap the native method, every llm has its own name for this
     # openai calls it create response, anthropic create message. so the name depends on your llm
     async def mynativemethod(self, *args, **kwargs):
-        
+
         # some details to get right here...
         # ensure conversation history is maintained. typically by passing it ie:
         if self._instructions:
             kwargs["system"] = [{"text": self._instructions}]
-            
+
         response_iterator = await self.client.mynativemethod(self, *args, **kwargs)
-        
+
         # while receiving streaming do this
         total_text = ""
         for chunk in response_iterator:
@@ -39,7 +39,7 @@ class MyLLM(LLM):
                     delta=chunk.text,
                 ))
             total_text += chunk.text
-            
+
         llm_response = LLMResponseEvent(response_iterator, total_text)
         # and when completed
         self.events.send(LLMResponseCompletedEvent(
@@ -57,7 +57,7 @@ class MyLLM(LLM):
         # call the LLM with the given text
         # be sure to use the streaming version
         self.mynativemethod(...)
-    
+
     @staticmethod
     def _normalize_message(my_input) -> List["Message"]:
         # convert the message to a list of messages so our conversation storage gets it
@@ -77,4 +77,4 @@ class MyLLM(LLM):
 If you need more examples look in
 
 - gemini_llm.py
-- aws_llm.py (AWS Bedrock implementation)
+- aws_llm.py (AWS Bedrock implementation)