"""Utilities for execution-feedback code generation experiments."""

from .agent import AgentResult, AttemptRecord, ExecutionFeedbackAgent
from .backends import OpenAICompatibleGenerator
from .dataset import CodeRLDataset, collate_code_rl
from .reward import CodeReward, score_code_batch, score_verification
from .evaluation import (
    EvaluationReport,
    aggregate_reports,
    estimate_pass_at_k,
    evaluate_candidates,
)
from .verifier import (
    CodeTask,
    TestCase,
    VerificationResult,
    VerificationStatus,
    extract_python_code,
    verify_code,
)

__all__ = [
    "AgentResult",
    "AttemptRecord",
    "CodeReward",
    "CodeRLDataset",
    "CodeTask",
    "EvaluationReport",
    "ExecutionFeedbackAgent",
    "OpenAICompatibleGenerator",
    "TestCase",
    "VerificationResult",
    "VerificationStatus",
    "aggregate_reports",
    "collate_code_rl",
    "estimate_pass_at_k",
    "evaluate_candidates",
    "extract_python_code",
    "score_verification",
    "score_code_batch",
    "verify_code",
]