Skip to content

[GSK-4033] Fix correctness aggregation error #2090

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

Merged
kevinmessiaen merged 4 commits into from
Jan 30, 2025
Merged

[GSK-4033] Fix correctness aggregation error #2090

Changes from 3 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
481f91d
Raise exception if correctness is string
henchaves Dec 13, 2024
2ce7fca
Update the condition to raise LLMGenerationError
henchaves Dec 19, 2024
effa6a0
Merge branch 'main' into feature/gsk-4033-fix-correctness-aggregation…
henchaves Dec 19, 2024
abeb48d
Merge branch 'main' into feature/gsk-4033-fix-correctness-aggregation…
henchaves Jan 29, 2025
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
10 changes: 9 additions & 1 deletion giskard/rag/metrics/correctness.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,13 +109,21 @@ def __call__(self, question_sample: dict, answer: AgentAnswer) -> dict:
temperature=0,
format="json_object",
)
return parse_json_output(

json_output = parse_json_output(
out.content,
llm_client=llm_client,
keys=["correctness", "correctness_reason"],
caller_id=self.__class__.__name__,
)

if "correctness" in json_output and not isinstance(json_output["correctness"], bool):
raise LLMGenerationError(
f"Error in correctness evaluation: {json_output['correctness']}. Please make sure the agent answer is correctly formatted."
Copy link
Copy Markdown
Member

@kevinmessiaen kevinmessiaen Jan 7, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Switching to repr is better so we can differentiate "True" from True, also

Suggested change
f"Error in correctness evaluation: {json_output['correctness']}. Please make sure the agent answer is correctly formatted."
f"Error in correctness evaluation: {repr(json_output['correctness'])}. Please make sure the agent answer is correctly formatted. Correctness value should be of boolean type."

)

return json_output

except Exception as err:
raise LLMGenerationError("Error while evaluating the agent") from err

Expand Down