LangChain 1.5.1 Types of output parser

침조 : https://python.langchain.com/docs/modules/model_io/output_parsers/

Output Parsers | 🦜️🔗 Langchain

 

 

CSV parser

from langchain.output_parsers import CommaSeparatedListOutputParser
from langchain.prompts import PromptTemplate
from langchain_openai import ChatOpenAI

output_parser = CommaSeparatedListOutputParser()

format_instructions = output_parser.get_format_instructions()
prompt = PromptTemplate(
    template="List five {subject}.\n{format_instructions}",
    input_variables=["subject"],
    partial_variables={"format_instructions": format_instructions},
)

model = ChatOpenAI(temperature=0)

chain = prompt | model | output_parser

result = chain.invoke({"subject": "ice cream flavors"})
print('Output>', result)

print('Output>')
for s in chain.stream({"subject": "ice cream flavors"}):
    print(s)

Output> ['Vanilla', 'Chocolate', 'Strawberry', 'Mint Chocolate Chip', 'Cookies and Cream']
Output>
['Vanilla']
['Chocolate']
['Strawberry']
['Mint Chocolate Chip']
['Cookies and Cream']

 

 

Datetime parser

from langchain.output_parsers import DatetimeOutputParser
from langchain.prompts import PromptTemplate
from langchain_openai import OpenAI

output_parser = DatetimeOutputParser()
template = """Answer the users question:

{question}

{format_instructions}"""
prompt = PromptTemplate.from_template(
    template,
    partial_variables={"format_instructions": output_parser.get_format_instructions()},
)

print('Output>', prompt)

chain = prompt | OpenAI() | output_parser

output = chain.invoke({"question": "when was bitcoin founded?"})

print('Output>', output)

Output> input_variables=['question'] partial_variables={'format_instructions': "Write a datetime string that matches the following pattern: '%Y-%m-%dT%H:%M:%S.%fZ'.\n\nExamples: 1725-03-05T17:26:45.931549Z, 854-01-11T20:07:28.505493Z, 1304-03-29T21:09:39.872406Z\n\nReturn ONLY this string, no other words!"} template='Answer the users question:\n\n{question}\n\n{format_instructions}'
Output> 2009-01-03 18:15:05

 

 

Json parser

from typing import List

from langchain.prompts import PromptTemplate
from langchain_core.output_parsers import JsonOutputParser
from langchain_core.pydantic_v1 import BaseModel, Field
from langchain_openai import ChatOpenAI

model = ChatOpenAI(temperature=0)

# Define your desired data structure.
class Joke(BaseModel):
    setup: str = Field(description="question to set up a joke")
    punchline: str = Field(description="answer to resolve the joke")


# And a query intented to prompt a language model to populate the data structure.
joke_query = "Tell me a joke."

# Set up a parser + inject instructions into the prompt template.
parser = JsonOutputParser(pydantic_object=Joke)

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

chain = prompt | model | parser

result = chain.invoke({"query": joke_query})
print('Output>', result)

Output> {'setup': "Why couldn't the bicycle stand up by itself?", 'punchline': 'Because it was two tired!'}

 

 

OpenAI Functions

from langchain_community.utils.openai_functions import (
    convert_pydantic_to_openai_function,
)
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.pydantic_v1 import BaseModel, Field, validator
from langchain_openai import ChatOpenAI

class Joke(BaseModel):
    """Joke to tell user."""

    setup: str = Field(description="question to set up a joke")
    punchline: str = Field(description="answer to resolve the joke")


openai_functions = [convert_pydantic_to_openai_function(Joke)]

model = ChatOpenAI(temperature=0)

prompt = ChatPromptTemplate.from_messages(
    [("system", "You are helpful assistant"), ("user", "{input}")]
)

from langchain.output_parsers.openai_functions import JsonOutputFunctionsParser

parser = JsonOutputFunctionsParser()
chain = prompt | model.bind(functions=openai_functions) | parser

result = chain.invoke({"input": "tell me a joke"})

print('Output>', result)

print('Output>')
for s in chain.stream({"input": "tell me a joke"}):
    print(s)

/DATA/venv/lib/python3.10/site-packages/langchain_core/_api/deprecation.py:117: LangChainDeprecationWarning: The function `convert_pydantic_to_openai_function` was deprecated in LangChain 0.1.16 and will be removed in 0.2.0. Use langchain_core.utils.function_calling.convert_to_openai_function() instead.
  warn_deprecated(
Output> {'setup': "Why couldn't the bicycle stand up by itself?", 'punchline': 'It was two tired!'}
Output>
{}
{'setup': ''}
{'setup': 'Why'}
{'setup': 'Why couldn'}
{'setup': "Why couldn't"}
{'setup': "Why couldn't the"}
{'setup': "Why couldn't the bicycle"}
{'setup': "Why couldn't the bicycle stand"}
{'setup': "Why couldn't the bicycle stand up"}
{'setup': "Why couldn't the bicycle stand up by"}
{'setup': "Why couldn't the bicycle stand up by itself"}
{'setup': "Why couldn't the bicycle stand up by itself?"}
{'setup': "Why couldn't the bicycle stand up by itself?", 'punchline': ''}
{'setup': "Why couldn't the bicycle stand up by itself?", 'punchline': 'It'}
{'setup': "Why couldn't the bicycle stand up by itself?", 'punchline': 'It was'}
{'setup': "Why couldn't the bicycle stand up by itself?", 'punchline': 'It was two'}
{'setup': "Why couldn't the bicycle stand up by itself?", 'punchline': 'It was two tired'}
{'setup': "Why couldn't the bicycle stand up by itself?", 'punchline': 'It was two tired!'}