Spaces:
Sleeping
Sleeping
File size: 1,899 Bytes
b7e6f83 | 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 | from __future__ import annotations
from dataclasses import dataclass
from typing import Any
from persistentpoker_bench.adapters.litellm_adapter import (
LiteLLMConfig,
extract_usage_summary,
request_decision_via_litellm,
)
from persistentpoker_bench.hand_runner import DecisionAgent, DecisionEnvelope
from persistentpoker_bench.prompting import PromptBundle
from persistentpoker_bench.retries import RetryPolicy
@dataclass(frozen=True, slots=True)
class RuntimeDecisionEnvelope(DecisionEnvelope):
provider: str
model_id: str
latency_seconds: float
usage: dict[str, Any]
@dataclass(slots=True)
class LiteLLMRuntimeAgent(DecisionAgent):
provider: str
config: LiteLLMConfig
retry_policy: RetryPolicy | None = None
def decide(
self,
*,
prompt_bundle: PromptBundle,
game_snapshot: dict[str, Any],
legal_actions_snapshot: dict[str, Any],
player_index: int,
hand_state: Any,
persistent_pool: Any,
) -> RuntimeDecisionEnvelope:
result = request_decision_via_litellm(
prompt_bundle=prompt_bundle,
config=self.config,
retry_policy=self.retry_policy,
)
usage = extract_usage_summary(result.raw_response)
return RuntimeDecisionEnvelope(
decision=result.decision,
raw_text=result.raw_text,
parse_mode=result.parse_mode,
attempts=result.attempts,
provider=self.provider,
model_id=self.config.model,
latency_seconds=result.latency_seconds,
usage=usage,
)
def runtime_envelope_to_dict(envelope: RuntimeDecisionEnvelope) -> dict[str, Any]:
return {
"provider": envelope.provider,
"model_id": envelope.model_id,
"latency_seconds": envelope.latency_seconds,
"usage": envelope.usage,
}
|