Kit Rhett Aultman 6e8b3d9e8e llm_client_backend: add object type to completion marker ...

The OpenAI chat completion API shows that streaming chat completion
chunks should have the object type "chat.completion.chunk" in the
completion marker chunk.  Without this, some parsers of the OpenAI API
don't handle chat completion streaming responses from this server.

This commit fixes the lack of an object type, keeps compliance with the
described API in the OpenAI docs (see
https://platform.openai.com/docs/api-reference/chat/create), and ensures
parsers like the Home Assistant HomeLLM addon work with it.

2026-01-18 14:09:58 -05:00

..

__init__.py

[clean] Clean up the code

2025-04-08 18:07:54 +08:00

asr_client_backend.py

[fix] Fix language parameter passing

2025-04-17 18:28:17 +08:00

base_model_backend.py

[perf] Optimize audio generation iterator and clean up code

2025-02-24 15:38:19 +08:00

chat_schemas.py

[perf] Optimize code directory structure

2025-02-20 17:18:20 +08:00

llm_client_backend.py

llm_client_backend: add object type to completion marker

2026-01-18 14:09:58 -05:00

openai_proxy_backend.py

[perf] Optimize audio generation iterator and clean up code

2025-02-24 15:38:19 +08:00

tts_client_backend.py

[feat] The tts backend supports model switching.

2025-04-23 14:47:10 +08:00

vision_model_backend.py

[perf] Optimize audio generation iterator and clean up code

2025-02-24 15:38:19 +08:00