Source code for aymara_ai.types.eval_analyze_response
# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
from typing import List, Optional
from datetime import datetime
from .eval import Eval
from .._models import BaseModel
__all__ = ["EvalAnalyzeResponse", "Result", "ResultStats"]
[docs]
class ResultStats(BaseModel):
total_responses_scored: int
"""Total number of responses scored across all runs"""
total_score_runs: int
"""Total number of score runs for this eval"""
avg_pass_rate: Optional[float] = None
"""Average pass rate across all score runs"""
best_pass_rate: Optional[float] = None
"""Best (highest) pass rate achieved"""
last_run_date: Optional[datetime] = None
"""Date of the most recent score run"""
worst_pass_rate: Optional[float] = None
"""Worst (lowest) pass rate achieved"""
[docs]
class Result(BaseModel):
eval: Eval
"""The eval data"""
stats: ResultStats
"""Aggregated statistics for this eval"""
[docs]
class EvalAnalyzeResponse(BaseModel):
has_more: bool
"""Whether there are more results available"""
results: List[Result]
"""List of matching evals with statistics"""
total_count: int
"""Total number of evals matching the analysis criteria"""