LangChain LCEL 链式调用

# 准备大模型实例

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
import os

os.environ["OPENAI_API_KEY"] = 'moonshot api key'
os.environ["OPENAI_API_BASE"] = 'https://api.moonshot.cn/v1/'

from langchain_openai import ChatOpenAI

api_key = os.getenv("OPENAI_API_KEY")
base_url = os.getenv("OPENAI_API_BASE")

print(api_key, base_url)

model = ChatOpenAI(
    openai_api_base=base_url,
    openai_api_key=api_key,
    model_name="moonshot-v1-8k",
    temperature=1,
)

# 案例 1

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
from operator import itemgetter

from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import ChatPromptTemplate

prompt1 = ChatPromptTemplate.from_template("what is the city {person} is from?")
prompt2 = ChatPromptTemplate.from_template(
    "what country is the city {city} in? respond in {language}"
)

chain1 = prompt1 | model | StrOutputParser()

chain2 = (
        {"city": chain1, "language": itemgetter("language")}
        | prompt2
        | model
        | StrOutputParser()
)

# 调用 1

1
chain2.invoke({"person": "小美", "language": "中文"})

‘根据对话中提供的信息，小美来自中国。具体来说，她提到她来自中国的 “大城市”。由于对话中没有明确提及，所以不清楚她具体来自哪个城市。’

# 案例 2

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
from langchain_core.runnables import RunnablePassthrough

prompt1 = ChatPromptTemplate.from_template(
    "generate a {attribute} color. Return the name of the color and nothing else:"
)
prompt2 = ChatPromptTemplate.from_template(
    "what is a fruit of color: {color}. Return the name of the fruit and nothing else:"
)
prompt3 = ChatPromptTemplate.from_template(
    "what is a country with a flag that has the color: {color}. Return the name of the country and nothing else:"
)
prompt4 = ChatPromptTemplate.from_template(
    "What is the color of {fruit} and the flag of {country}?"
)

model_parser = model | StrOutputParser()

color_generator = (
        {"attribute": RunnablePassthrough()} | prompt1 | {"color": model_parser}
)
color_to_fruit = prompt2 | model_parser
color_to_country = prompt3 | model_parser
question_generator = (
        color_generator | {"fruit": color_to_fruit, "country": color_to_country} | prompt4
)

# 调用 1

1
question_generator.invoke("warm")

ChatPromptValue(messages=[HumanMessage(content=‘What is the color of Peach and the flag of Georgia?’)])

# 调用 2

1
2
prompt = question_generator.invoke("warm")
model.invoke(prompt)

AIMessage(content=“The color of a peach can vary depending on its ripeness. Generally, a ripe peach is characterized
by a warm, yellow-orange hue with some variations of red or pink shades on its skin. These colors give the fruit an
appealing and vibrant appearance.\n\nThe flag of Georgia, referring to the U.S. state and not to the country located
at
the intersection of Europe and Asia, consists of three equal horizontal bands. The colors of the bands are as follows:
\n\n1. Top band: Dark red - this represents the courage and determination of the people of Georgia.\n2. Middle band:
White - this symbolizes the mountain peaks and the virtue of purity in the state.\n3. Bottom band: Sky blue - this
signifies the Georgia state’s hospitality and staying power.\n\nIn summary, the color of a ripe peach typically
consists
of yellow-orange, red, or pink shades, while the flag of Georgia, the U.S. state, comprises dark red, white, and sky
blue bands.”, response_metadata={‘token_usage’: {‘completion_tokens’: 198, ‘prompt_tokens’: 15, ‘total_tokens’:
213}, ’
model_name’: ‘moonshot-v1-8k’, ‘system_fingerprint’: None, ‘finish_reason’: ‘stop’, ‘logprobs’: None})

# 案例 3

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
planner = (
        ChatPromptTemplate.from_template("生成关于: {input}的论点")
        | model
        | StrOutputParser()
        | {"base_response": RunnablePassthrough()}
)

arguments_for = (
        ChatPromptTemplate.from_template(
            "列出 {base_response} 的优点或积极方面"
        )
        | model
        | StrOutputParser()
)
arguments_against = (
        ChatPromptTemplate.from_template(
            "列出 {base_response} 的缺点或负面方面"
        )
        | model
        | StrOutputParser()
)

final_responder = (
        ChatPromptTemplate.from_messages(
            [
                ("ai", "{original_response}"),
                ("human", "Pros:\n{results_1}\n\nCons:\n{results_2}"),
                ("system", "根据批评生成一个最终回应"),
            ]
        )
        | model
        | StrOutputParser()
)

chain = (
        planner
        | {
            "results_1": arguments_for,
            "results_2": arguments_against,
            "original_response": itemgetter("base_response"),
        }
        | final_responder
)

# itemgetter 作为从映射中提取数据的简写

chain.invoke({"input": "内卷"})

内卷现象作为一个复杂的社会现象，既包含了一定的积极作用，也带来了许多负面影响。在应对内卷现象时，我们应该全面地认识其正反两面，采取有力的措施来平衡这些利弊，引导社会走向更加健康和谐的发展道路。
首先，我们应该正视内卷现象带来的心理健康问题、人际关系紧张、工作与生活失衡等问题。政府、企业和教育机构需要共同努力，通过制定合理的政策、关注员工福利、培养学生的创新能力和批判性思维等措施，来减轻内卷现象带来的负面影响。
同时，我们也应该看到内卷现象在一定程度上激发了人们的勤奋努力、提高竞争力和创新能力。因此，在应对内卷现象时，我们应该鼓励人们充分发挥自己的特长，关注个体差异，从而推动社会的多元化发展。
此外，社会各界应该共同努力提高人们的心理素质，关注心理健康问题，帮助人们在面对竞争和压力时保持良好的心态。同时，我们需要提高人们对教育、工作和生活品质的认识，使人们在追求物质生活和职业成功的同时，不忘关注家庭、友情等生活的重要方面。
总之，内卷现象作为一个具有双重性的社会现象，我们应该在全面认识其利弊的基础上，采取有效的措施来应对挑战，促进社会的平衡、健康和可持续发展。