Skip to content

fix(wren-ai-service): handle json_schema response_format for non-OpenAI models and guard empty SQL diagnosis response #2167

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Thangoun wants to merge 3 commits into
base: main
Choose a base branch
from
Open

fix(wren-ai-service): handle json_schema response_format for non-OpenAI models and guard empty SQL diagnosis response #2167

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
5dfdafc
fix(wren-ai-service): strip json_schema response_format for non-OpenA…
Mar 24, 2026
8c5e239
fix(wren-ai-service): address review comments and improve json_schema…
Mar 24, 2026
7be1c2d
Merge branch 'main' into fix/json-schema-empty-response
Thangoun Mar 27, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 5 additions & 2 deletions wren-ai-service/src/pipelines/generation/sql_diagnosis.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -91,8 +91,11 @@ async def generate_sql_diagnosis(
@observe(capture_input=False)
async def post_process(
generate_sql_diagnosis: dict,
) -> str:
return orjson.loads(generate_sql_diagnosis.get("replies")[0])
) -> dict:
reply = generate_sql_diagnosis.get("replies", [""])[0]
if not reply or not reply.strip():
return {"reasoning": "LLM did not return any response."}
return orjson.loads(reply)


## End of Pipeline
Expand Down
10 changes: 9 additions & 1 deletion wren-ai-service/src/providers/llm/litellm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -101,8 +101,16 @@ async def _run(

generation_kwargs = {
**combined_generation_kwargs,
**(generation_kwargs or {}),
**(generation_kwargs or {}),
}
# Strip response_format with type=json_schema — only supported by OpenAI
# native models. Custom api_base models (e.g. Ollama, LiteLLM proxy with
# non-OpenAI backends) return empty responses when this is forwarded.
# System prompts already include explicit JSON format instructions.
if self._api_base and isinstance(
generation_kwargs.get("response_format"), dict
) and generation_kwargs["response_format"].get("type") == "json_schema":
generation_kwargs.pop("response_format")

allowed_openai_params = generation_kwargs.get(
"allowed_openai_params", []
Expand Down