Spaces:
Sleeping
Sleeping
File size: 2,562 Bytes
71cb877 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 | # Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the BSD-style license found in the
# LICENSE file in the root directory of this source tree.
"""AI Response Evaluation Environment Client."""
from typing import Dict
from openenv.core import EnvClient
from openenv.core.client_types import StepResult
from openenv.core.env_server.types import State
from .models import CodeAssessmentAction, CodeAssessmentObservation
class CodeAssessmentEnv(
EnvClient[CodeAssessmentAction, CodeAssessmentObservation, State]
):
"""
Client for the AI Response Evaluation Environment.
Example:
>>> env = await CodeAssessmentEnv.from_docker_image("code_assessment_env:latest")
>>> result = await env.reset()
>>> print(result.observation.task_type)
>>> result = await env.step(CodeAssessmentAction(answer="incorrect, factual-error"))
"""
def _step_payload(self, action: CodeAssessmentAction) -> Dict:
return {"answer": action.answer}
def _parse_result(self, payload: Dict) -> StepResult[CodeAssessmentObservation]:
obs_data = payload.get("observation", {})
observation = CodeAssessmentObservation(
problem_description=obs_data.get("problem_description", ""),
difficulty=obs_data.get("difficulty", "easy"),
test_case_input=obs_data.get("test_case_input", ""),
task_type=obs_data.get("task_type", "correctness_check"),
language=obs_data.get("language", "en"),
user_age=obs_data.get("user_age"),
user_mood=obs_data.get("user_mood"),
user_context=obs_data.get("user_context"),
expected_output=obs_data.get("expected_output"),
feedback=obs_data.get("feedback", ""),
is_correct=obs_data.get("is_correct", False),
partial_credit=obs_data.get("partial_credit", 0.0),
problems_solved=obs_data.get("problems_solved", 0),
current_streak=obs_data.get("current_streak", 0),
done=payload.get("done", False),
reward=payload.get("reward"),
metadata=obs_data.get("metadata", {}),
)
return StepResult(
observation=observation,
reward=payload.get("reward"),
done=payload.get("done", False),
)
def _parse_state(self, payload: Dict) -> State:
return State(
episode_id=payload.get("episode_id"),
step_count=payload.get("step_count", 0),
)
|