07. EnumOutputParser

EnumOutputParser

Let's find out how to use the Enum output parser!

from langchain.output_parsers.enum import EnumOutputParser

• enum Using module Colors Define the class.

• Colors Class Enum Inheritance, RED , GREEN , BLUE It has three color values.

from enum import Enum


class Colors(Enum):
    RED = "RED"
    GREEN = "GREEN"
    BLUE = "BLUE"

# EnumOutputParser create immediately
parser = EnumOutputParser(enum=Colors)

• People's information at the prompt ( {person} ) And parsing instructions ( {instructions} ).

• parser.get_format_instructions() Call the function to get the parsing instructions.

• prompt, ChatOpenAI The model, parser is connected to form a processing chain.

from langchain_core.prompts import PromptTemplate
from langchain_openai import ChatOpenAI

# Create a prompt template.
prompt = PromptTemplate.from_template(
    """What color is the following object?

Object: {object}

Instructions: {instructions}"""
    # Get the instruction format from the parser and apply it partially.
).partial(instructions=parser.get_format_instructions())
# Prompt and ChatOpenAI, Connect the parser.
chain = prompt | ChatOpenAI() | parser

• chain.invoke Use the function to request information about the "sky".

response = chain.invoke({"object": "sky"})  # "sky" Execute chain calls for
print(response)

Colors.BLUE

Check the results.