Fix Google GenAI streaming handler to use async generator

- Keep stream_generator as async function (not sync)
- Wrap Google's synchronous iterator in async generator
- Properly structure if/else for streaming vs non-streaming paths
- Fix 'client has been closed' error in streaming responses

This fixes the issue where streaming requests through autoselect
were failing with 'Cannot send a request, as a client has been closed'
error.

aisbf/providers.py

@@ -160,23 +160,14 @@ class GoogleProviderHandler(BaseProviderHandler):
                 )
                 logging.info(f"GoogleProviderHandler: Streaming response received")
                 self.record_success()
-                return response
-            else:
-                # Generate content using the google-genai client
-                response = self.client.models.generate_content(
-                    model=model,
-                    contents=content,
-                    config=config
-                )
                 
-                # Handle streaming response
-                if stream:
-                    logging.info(f"GoogleProviderHandler: Processing streaming response")
-                    
-                    # Create a generator that yields OpenAI-compatible chunks
+                # Create an async generator that yields OpenAI-compatible chunks
+                # Google's generate_content_stream() returns a synchronous iterator
+                # We need to wrap it in an async generator
                 async def stream_generator():
                     try:
                         chunk_id = 0
+                        # Iterate over the sync iterator
                         for chunk in response:
                             logging.info(f"GoogleProviderHandler: Processing stream chunk")
                             
@@ -210,7 +201,7 @@ class GoogleProviderHandler(BaseProviderHandler):
                             
                             chunk_id += 1
                             logging.info(f"Yielding OpenAI chunk: {openai_chunk}")
-                                # Yield the complete chunk object as a single line
+                            # Yield to complete chunk object as a single line
                             yield openai_chunk
                             
                     except Exception as e:
@@ -218,8 +209,15 @@ class GoogleProviderHandler(BaseProviderHandler):
                         raise
                 
                 return stream_generator()
+            else:
+                # Non-streaming request
+                # Generate content using the google-genai client
+                response = self.client.models.generate_content(
+                    model=model,
+                    contents=content,
+                    config=config
+                )
 
-                # Non-streaming response
                 logging.info(f"GoogleProviderHandler: Response received: {response}")
                 self.record_success()