Skip to content

fix/add-auto-retry #19

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
fengzz-coding wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix/add-auto-retry #19

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
16 changes: 14 additions & 2 deletions validator/modules/llm_judge/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -509,7 +509,7 @@ def _call_gpt(
def log_retry(retry_state):
logger.warning(
f"API call failed (attempt {retry_state.attempt_number}/3), "
f"retrying in {retry_state.next_action.sleep:.1f}s: {retry_state.outcome.exception()}"
f"retrying in {retry_state.next_action.sleep:.1f}s: {str(retry_state.outcome.exception())[:200]}"
)

@retry(
Expand All @@ -520,6 +520,7 @@ def log_retry(retry_state):
)
def _make_api_call():
completion = self.client.chat.completions.create(**params)
self._parse_check_valid_json(completion.choices[0].message.content)
return completion.choices[0].message.content

try:
Expand Down Expand Up @@ -821,6 +822,17 @@ def _construct_evaluation_prompt(
{"role": "system", "content": system_prompt},
]

def _parse_check_valid_json(self, response: str,):
json_match = re.search(r'\{[^}]*"score"[^}]*\}', response, re.DOTALL)
if json_match:
json_str = json_match.group(0)
parsed_json = json.loads(json_str)
_ = float(parsed_json["score"])
_ = float(parsed_json["confidence"])
_ = str(parsed_json["reasoning"])
else:
raise

def _parse_llm_response(
self, response: str, model_name: str = None
) -> Dict[str, Any]:
Expand Down Expand Up @@ -988,7 +1000,7 @@ def validate(self, data: LLMJudgeInputData, **kwargs) -> LLMJudgeMetrics:
f"Evaluation progress: {completed_count}/{total_tasks} conversations evaluated ({evaluations_done}/{total_eval_calls} LLM calls, {progress_pct:.1f}%)"
)
except Exception as e:
logger.error(f"Evaluation task failed: {e}")
logger.error(f"Evaluation task failed: {str(e)[:200]}")
completed_count += 1
# Add default result for failed tasks
evaluation_results.append(
Expand Down