通义千问版:基于LangChain的LLM应用开发2——模型、提示和输出解析


参考文档

https://cloud.baidu.com/qianfandev/topic/267899

https://python.langchain.com/v0.1/docs/integrations/llms/tongyi/

运行环境

本文档的内容是 .ipynb 文件格式,因此可以在 Google Colab 上运行。

当然也可以在安装了 Jupter 插件的 VSCode 下运行。

因为本文使用的是国内通义千问的模型,因此不用科学上网也可以执行。

源码

https://gitee.com/tanggaowei/langchain-llm-app-dev

安装相关模块

%pip install python-dotenv
%pip install --upgrade langchain
%pip install --upgrade langchain_community
%pip install --upgrade --quiet dashscope

加载环境变量

from pathlib import Path
from dotenv import load_dotenv, find_dotenv

# 加载环境变量,每次都覆盖旧数据

env_path = Path(‘.’) / ‘test.env’

load_dotenv(dotenv_path=env_path, override=True)

test.env 中保存了环境变量的值:

DASHSCOPE_API_KEY=sk-***

初始化大语言模型对象

import os
from langchain_community.llms import Tongyi


Tongyi().invoke(“哪支NFL球队在贾斯汀·比伯出生的那一年赢得了超级碗?”)

使用大模型翻译文本

customer_email = """
Arrr, I be fuming that me blender lid \
flew off and splattered me kitchen walls \
with smoothie! And to make matters worse,\
the warranty don't cover the cost of \
cleaning up me kitchen. I need yer help \
right now, matey!
"""

language = “chinese”

prompt = f”””Translate the text \

into {language}.

text: “`{customer_email}“`

“””

print(prompt)

response = Tongyi().invoke(prompt)

print(response)

通过LangChain的方式来翻译文本

from langchain.prompts import PromptTemplate

llm = Tongyi()

template = “””Translate the text \

into {language}. \

text: “`{text}“`

“””

prompt = PromptTemplate.from_template(template)

chain = prompt | llm

language = “chinese”

text = “””

Arrr, I be fuming that me blender lid \

flew off and splattered me kitchen walls \

with smoothie! And to make matters worse, \

the warranty don’t cover the cost of \

cleaning up me kitchen. I need yer help \

right now, matey!

“””

response = chain.invoke({“language”: “chinese”, “text”: text})

print(response)

response = chain.invoke({“language”: “janpnese”, “text”: text})

print(response)

  

通过解析器来解析模型输出

解析器的作用就是将一定格式的文本(例如json文本),解析成python对象(例如dict类型对象),以便进一步处理。

注意:解析器并不是格式化文本的工具,它是将已经格式化的文本转化为python对象。

customer_review = """\
This leaf blower is pretty amazing. It has four settings:\
candle blower, gentle breeze, windy city, and tornado. \
It arrived in two days, just in time for my wife's \
anniversary present. \
I think my wife liked it so much she was speechless. \
So far I've been the only one using it, and I've been \
using it every other morning to clear the leaves on our lawn. \
It's slightly more expensive than the other leaf blowers \
out there, but I think it's worth it for the extra features.
"""


review_template = “””\

For the following text, extract the following information:

gift: Was the item purchased as a gift for someone else? \

Answer true if yes, false if not or unknown.

delivery_days: How many days did it take for the product \

to arrive? If this information is not found, output -1.

price_value: Extract any sentences about the value or price,\

and output them as a comma separated Python list.

Format the output as JSON with the following keys, without ““`”:

gift

delivery_days

price_value

text: {text}

“””

from langchain.prompts import ChatPromptTemplate

prompt = ChatPromptTemplate.from_template(review_template)

chain = prompt | llm

response = chain.invoke({“text”:customer_review})

print(response)

from langchain.output_parsers import ResponseSchema, StructuredOutputParser

# description 相当于字段的描述,不影响解析结果

gift_schema = ResponseSchema(name=”gift”,

type=”bool”,

description=”是否为礼物”)

delivery_days_schema = ResponseSchema(name=”delivery_days”,

type=”string”,

description=”几天到货”)

price_value_schema = ResponseSchema(name=”price_value”,

type=”string”,

description=”价值或价格内容”)

response_schemas = [gift_schema,delivery_days_schema,price_value_schema]

output_parser = StructuredOutputParser.from_response_schemas(response_schemas)

format_instructions = output_parser.get_format_instructions()

print(format_instructions)

prompt = ChatPromptTemplate.from_template(template=review_template,format_instructions=format_instructions)

chain = prompt | llm

response = chain.invoke({“text”:customer_review})

print(response)

# 将字符串格式化为 dict 类型变量

output_dict = output_parser.parse(response)

print(output_dict)

print(output_dict.get(‘gift’))


前一篇:
后一篇:

发表评论