fix: resolve streaming duplication and markdown formatting issues

- Fix text duplication bug with proper delta calculation
- Preserve markdown formatting with chunk-based delivery (50 chars)
- Handle GeneratorExit errors from async context managers
- Update Chat service streaming to preserve formatting
- Ensure proper word-by-word streaming without duplicates

src/chat/service.py

@@ -214,9 +214,12 @@ async def create_chat_completion_stream(
             in_reasoning = False
 
         elif item.type == "message":
-            # Stream message content word by word
+            # Stream message content in chunks (preserves newlines, markdown, etc.)
             text = item.data["content"][0]["text"]
-            for word in text.split():
+            chunk_size = 50  # characters per chunk
+
+            for i in range(0, len(text), chunk_size):
+                chunk = text[i:i+chunk_size]
                 yield ChatCompletionChunk(
                     id=completion_id,
                     object=constants.CHAT_COMPLETION_CHUNK_OBJECT,
@@ -225,12 +228,12 @@ async def create_chat_completion_stream(
                     choices=[
                         ChatCompletionChunkChoice(
                             index=0,
-                            delta=ChatCompletionChunkDelta(content=f"{word} "),
+                            delta=ChatCompletionChunkDelta(content=chunk),
                             finish_reason=None,
                         )
                     ],
                 )
-                await asyncio.sleep(0.05)  # Simulate typing
+                await asyncio.sleep(0.02)  # Faster since chunks are larger
 
     # Final chunk with finish_reason
     yield ChatCompletionChunk(