MCPcopy
hub / github.com/algorithmicsuperintelligence/optillm / generate_streaming_response

Function generate_streaming_response

optillm/server.py:551–578  ·  view source on GitHub ↗
(final_response, model)

Source from the content-addressed store, hash-verified

549 return responses, total_tokens
550
551def generate_streaming_response(final_response, model):
552 # Generate a unique response ID
553 response_id = f"chatcmpl-{int(time.time()*1000)}"
554 created = int(time.time())
555
556 # Yield the final response with OpenAI-compatible format
557 if isinstance(final_response, list):
558 for index, response in enumerate(final_response):
559 # First chunk includes role
560 yield "data: " + json.dumps({
561 "id": response_id,
562 "object": "chat.completion.chunk",
563 "created": created,
564 "model": model,
565 "choices": [{"delta": {"role": "assistant", "content": response}, "index": index, "finish_reason": "stop"}],
566 }) + "\n\n"
567 else:
568 # First chunk includes role
569 yield "data: " + json.dumps({
570 "id": response_id,
571 "object": "chat.completion.chunk",
572 "created": created,
573 "model": model,
574 "choices": [{"delta": {"role": "assistant", "content": final_response}, "index": 0, "finish_reason": "stop"}],
575 }) + "\n\n"
576
577 # Yield the final message to indicate the stream has ended
578 yield "data: [DONE]\n\n"
579
580def extract_contents(response_obj):
581 contents = []

Callers 1

proxyFunction · 0.85

Calls

no outgoing calls

Tested by

no test coverage detected