當解析錯誤發生時，如何重試

雖然在某些情況下，僅查看輸出即可修正任何解析錯誤，但在其他情況下則不然。一個例子是輸出不僅格式不正確，而且部分不完整。請考慮以下範例。

from langchain.output_parsers import OutputFixingParser
from langchain_core.exceptions import OutputParserException
from langchain_core.output_parsers import PydanticOutputParser
from langchain_core.prompts import PromptTemplate
from langchain_openai import ChatOpenAI, OpenAI
from pydantic import BaseModel, Field

template = """Based on the user question, provide an Action and Action Input for what step should be taken.
{format_instructions}
Question: {query}
Response:"""


class Action(BaseModel):
    action: str = Field(description="action to take")
    action_input: str = Field(description="input to the action")


parser = PydanticOutputParser(pydantic_object=Action)

prompt = PromptTemplate(
    template="Answer the user query.\n{format_instructions}\n{query}\n",
    input_variables=["query"],
    partial_variables={"format_instructions": parser.get_format_instructions()},
)

prompt_value = prompt.format_prompt(query="who is leo di caprios gf?")

bad_response = '{"action": "search"}'

如果我們嘗試按原樣解析此回應，則會收到錯誤

try:
    parser.parse(bad_response)
except OutputParserException as e:
    print(e)

Failed to parse Action from completion {"action": "search"}. Got: 1 validation error for Action
action_input
  Field required [type=missing, input_value={'action': 'search'}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.9/v/missing
For troubleshooting, visit: https://langchain-python.dev.org.tw/docs/troubleshooting/errors/OUTPUT_PARSING_FAILURE

如果我們嘗試使用 OutputFixingParser 來修正此錯誤，它會感到困惑 - 也就是說，它不知道實際要為動作輸入放入什麼。

fix_parser = OutputFixingParser.from_llm(parser=parser, llm=ChatOpenAI())

fix_parser.parse(bad_response)

Action(action='search', action_input='input')

相反地，我們可以使用 RetryOutputParser，它會傳入提示（以及原始輸出）以再次嘗試獲得更好的回應。

from langchain.output_parsers import RetryOutputParser

API 參考：RetryOutputParser

retry_parser = RetryOutputParser.from_llm(parser=parser, llm=OpenAI(temperature=0))

retry_parser.parse_with_prompt(bad_response, prompt_value)

Action(action='search', action_input='leo di caprio girlfriend')

我們也可以使用自訂鏈輕鬆新增 RetryOutputParser，自訂鏈會將原始 LLM/ChatModel 輸出轉換為更可行的格式。

from langchain_core.runnables import RunnableLambda, RunnableParallel

completion_chain = prompt | OpenAI(temperature=0)

main_chain = RunnableParallel(
    completion=completion_chain, prompt_value=prompt
) | RunnableLambda(lambda x: retry_parser.parse_with_prompt(**x))


main_chain.invoke({"query": "who is leo di caprios gf?"})

API 參考：RunnableLambda | RunnableParallel

Action(action='search', action_input='leo di caprio girlfriend')

尋找 RetryOutputParser 的 API 文件。

此頁面是否對您有幫助？