langchain.smith.evaluation.runner_utils.run_on_dataset

langchain.smith.evaluation.runner_utils.run_on_dataset(client: Optional[Client], dataset_name: str, llm_or_chain_factory: Union[Callable[[], Union[Chain, Runnable]], BaseLanguageModel, Callable[[dict], Any], Runnable, Chain], *, evaluation: Optional[RunEvalConfig] = None, dataset_version: Optional[Union[datetime, str]] = None, concurrency_level: int = 5, project_name: Optional[str] = None, project_metadata: Optional[Dict[str, Any]] = None, verbose: bool = False, revision_id: Optional[str] = None, **kwargs: Any) → Dict[str, Any]

在数据集上运行链或语言模型并将跟踪信息存储到指定的项目名称。

参数

• dataset_name (str) – 要运行链的数据集名称。

• llm_or_chain_factory (Union[Callable[[], Union[Chain, Runnable]], BaseLanguageModel, Callable[[dict], Any], Runnable, Chain]) – 运行在数据集上的语言模型或链构造器。链构造器用于允许对每个示例独立调用而不携带状态。

• evaluation (Optional[RunEvalConfig]) – 评估器在链的结果上运行的配置

• concurrency_level (int) – 同时运行的异步任务数量。

• project_name (Optional[str]) – 将跟踪的数据存储在项目中的名称。默认为 {dataset_name}-{chain class name}-{datetime}。

• project_metadata (Optional[Dict[str, Any]]) – 可选项目元数据，用于存储测试变体的信息。（提示版本、模型版本等。）

• client (Optional[Client]) – 用于访问数据集和记录反馈以及运行跟踪的 LangSmith 客户端。

• verbose (bool) – 是否打印进度。

• tags – 添加到项目每个运行中的标签。

• revision_id (Optional[str]) – 可选修订标识符，将该测试运行分配给以跟踪系统不同版本的性能。

• dataset_version (Optional[Union[datetime, str]]) –

• kwargs (Any) –

返回

包含运行项目名称和模型输出的字典。

返回类型

Dict[str, Any]

有关此函数的大多数情况下更快版本的异步版本，请参阅arun_on_dataset()。

示例

from langsmith import Client
from langchain_openai import ChatOpenAI
from langchain.chains import LLMChain
from langchain.smith import smith_eval.RunEvalConfig, run_on_dataset

# Chains may have memory. Passing in a constructor function lets the
# evaluation framework avoid cross-contamination between runs.
def construct_chain():
    llm = ChatOpenAI(temperature=0)
    chain = LLMChain.from_string(
        llm,
        "What's the answer to {your_input_key}"
    )
    return chain

# Load off-the-shelf evaluators via config or the EvaluatorType (string or enum)
evaluation_config = smith_eval.RunEvalConfig(
    evaluators=[
        "qa",  # "Correctness" against a reference answer
        "embedding_distance",
        smith_eval.RunEvalConfig.Criteria("helpfulness"),
        smith_eval.RunEvalConfig.Criteria({
            "fifth-grader-score": "Do you have to be smarter than a fifth grader to answer this question?"
        }),
    ]
)

client = Client()
run_on_dataset(
    client,
    dataset_name="<my_dataset_name>",
    llm_or_chain_factory=construct_chain,
    evaluation=evaluation_config,
)

您还可以通过从 StringEvaluator 或 LangSmith 的 RunEvaluator 类派生来自定义评估器。

from typing import Optional
from langchain.evaluation import StringEvaluator

class MyStringEvaluator(StringEvaluator):

    @property
    def requires_input(self) -> bool:
        return False

    @property
    def requires_reference(self) -> bool:
        return True

    @property
    def evaluation_name(self) -> str:
        return "exact_match"

    def _evaluate_strings(self, prediction, reference=None, input=None, **kwargs) -> dict:
        return {"score": prediction == reference}

evaluation_config = smith_eval.RunEvalConfig(
    custom_evaluators = [MyStringEvaluator()],
)

run_on_dataset(
    client,
    dataset_name="<my_dataset_name>",
    llm_or_chain_factory=construct_chain,
    evaluation=evaluation_config,
)