轻松上手的LangChain学习说明书

prompt = """Translate the text \
that is delimited by triple backticks \
into a style that is {style}. \
text: ```{text}```
"""
1.
2.
3.
4.
5.

from langchain.prompts import ChatPromptTemplate
prompt_template=ChatPromptTemplate.from_template(prompt)
print(prompt_template,'ChatPromptTemplate')
1.
2.
3.

#使用 Python f 字符串模板：
from langchain.prompts import PromptTemplate
fstring_template = """Tell me a {adjective} joke about {content}"""
prompt = PromptTemplate.from_template(fstring_template)
print(prompt.format(adjective="funny", content="chickens"))
# Output: Tell me a funny joke about chickens.
1.
2.
3.
4.
5.
6.

#使用 jinja2 模板：
from langchain.prompts import PromptTemplate
jinja2_template = "Tell me a {{ adjective }} joke about {{ content }}"
prompt = PromptTemplate.from_template(jinja2_template, template_format="jinja2")
print(prompt.format(adjective="funny", content="chickens"))
# Output: Tell me a funny joke about chickens.
1.
2.
3.
4.
5.
6.

from langchain.prompts import PromptTemplate #用于 PromptTemplate 为字符串提示创建模板。
#默认情况下， PromptTemplate 使用 Python 的 str.format 语法进行模板化;但是可以使用其他模板语法（例如， jinja2 ）
prompt_template = PromptTemplate.from_template("Tell me a {adjective} joke about {content}.")
print(prompt_template.format(adjective="funny", content="chickens"))
1.
2.
3.
4.

Tell me a funny joke about chickens.
1.

#ChatPromptTemplate.from_messages 接受各种消息表示形式。
template = ChatPromptTemplate.from_messages([
("system", "You are a helpful AI bot. Your name is {name}."),
("human", "Hello, how are you doing?"),
("ai", "I'm doing well, thanks!"),
("human", "{user_input}"),
])
messages = template.format_messages(
name="Bob",
user_input="What is your name?"
)
print(messages)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.

[('system', 'You are a helpful AI bot. Your name is Bob.'),
 ('human', 'Hello, how are you doing?'),
 ('ai', "I'm doing well, thanks!"),
 ('human', 'What is your name?')]
1.
2.
3.
4.

from langchain.prompts import PromptTemplate
prompt = PromptTemplate(template="{foo}{bar}", input_variables=["foo", "bar"])

# 可以使用 PromptTemplate.partial() 方法创建部分提示模板。
partial_prompt = prompt.partial(foo="foo")
print(partial_prompt.format(bar="baz"))

#也可以只使用分部变量初始化提示。
prompt = PromptTemplate(template="{foo}{bar}", input_variables=["bar"], partial_variables={"foo": "foo"})
print(prompt.format(bar="baz"))
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.

foobaz
foobaz
1.
2.

from datetime import datetime

def _get_datetime():
    now = datetime.now()
    return now.strftime("%m/%d/%Y, %H:%M:%S")

prompt = PromptTemplate(
      template="Tell me a {adjective} joke about the day {date}",
      input_variables=["adjective", "date"]
)
partial_prompt = prompt.partial(date=_get_datetime)
print(partial_prompt.format(adjective="funny"))

# 除上述方法，部分函数声明和普通的prompt一样，也可以直接用partial_variables去声明
prompt = PromptTemplate(
template="Tell me a {adjective} joke about the day {date}",
input_variables=["adjective"],
partial_variables={"date": _get_datetime})
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.

Tell me a funny joke about the day 12/08/2022, 16:25:30
1.

from langchain.prompts.pipeline import PipelinePromptTemplate
from langchain.prompts.prompt import PromptTemplate

full_template = """{introduction}
{example}
"""
full_prompt = PromptTemplate.from_template(full_template)

introduction_template = """You are impersonating Elon Musk."""
introduction_prompt = PromptTemplate.from_template(introduction_template)

example_template = """Here's an example of an interaction """
example_prompt = PromptTemplate.from_template(example_template)

input_prompts = [("introduction", introduction_prompt),
("example", example_prompt),]

pipeline_prompt = PipelinePromptTemplate(final_prompt=full_prompt, pipeline_prompts=input_prompts)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.

import inspect

# 该函数将返回给定其名称的函数的源代码。 inspect作用就是获取源代码
def get_source_code(function_name):
    # Get the source code of the function
    return inspect.getsource(function_name)

# 测试函数
def test():
    return 1 + 1

from langchain.prompts import StringPromptTemplate
from pydantic import BaseModel, validator

# 初始化字符串prompt
PROMPT = """\
提供一个函数名和源代码并给出函数的相应解释
函数名: {function_name}
源代码:
{source_code}
解释:
"""

class FunctionExplainerPromptTemplate(StringPromptTemplate, BaseModel):
    """一个自定义提示模板，以函数名作为输入，并格式化提示模板以提供函数的源代码。 """
    @validator("input_variables")
    def validate_input_variables(cls, v):
        """验证输入变量是否正确。"""
        if len(v) != 1 or "function_name" not in v:
            raise ValueError("函数名必须是唯一的输入变量。")
        return v

    def format(self, **kwargs) -> str:
        # 获取源代码
        source_code = get_source_code(kwargs["function_name"])
        # 源代码+名字提供给prompt
        prompt = PROMPT.format(
            function_name=kwargs["function_name"].__name__, source_code=source_code)
        return prompt

    def _prompt_type(self):
        return "function-explainer"
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.

#初始化prompt实例
fn_explainer = FunctionExplainerPromptTemplate(input_variables=["function_name"])

# 定义函数 test_add
def test_add():
    return 1 + 1

# Generate a prompt for the function "test_add"
prompt_1 = fn_explainer.format(function_name=test_add)
print(prompt_1)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.

from langchain.prompts.example_selector import SemanticSimilarityExampleSelector
from langchain.vectorstores import Chroma
from langchain.embeddings import OpenAIEmbeddings
from langchain.prompts import FewShotPromptTemplate, PromptTemplate

examples = [
    {"question": "Who lived longer, Muhammad Ali or Alan Turing?",
    "answer":
    """
    Are follow up questions needed here: Yes.
    Follow up: How old was Muhammad Ali when he died?
    Intermediate answer: Muhammad Ali was 74 years old when he died.
    Follow up: How old was Alan Turing when he died?
    Intermediate answer: Alan Turing was 41 years old when he died.
    So the final answer is: Muhammad Ali
    """},
    {"question": "When was the founder of craigslist born?",
    "answer":
    """
    Are follow up questions needed here: Yes.
    Follow up: Who was the founder of craigslist?
    Intermediate answer: Craigslist was founded by Craig Newmark.
    Follow up: When was Craig Newmark born?
    Intermediate answer: Craig Newmark was born on December 6, 1952.
    So the final answer is: December 6, 1952
    """},
    {"question": "Who was the maternal grandfather of George Washington?",
    "answer":
    """
    Are follow up questions needed here: Yes.
    Follow up: Who was the mother of George Washington?
    Intermediate answer: The mother of George Washington was Mary Ball Washington.
    Follow up: Who was the father of Mary Ball Washington?
    Intermediate answer: The father of Mary Ball Washington was Joseph Ball.
    So the final answer is: Joseph Ball
    """},
    {"question": "Are both the directors of Jaws and Casino Royale from the same country?",
    "answer":
    """
    Are follow up questions needed here: Yes.
    Follow up: Who is the director of Jaws?
    Intermediate Answer: The director of Jaws is Steven Spielberg.
    Follow up: Where is Steven Spielberg from?
    Intermediate Answer: The United States.
    Follow up: Who is the director of Casino Royale?
    Intermediate Answer: The director of Casino Royale is Martin Campbell.
    Follow up: Where is Martin Campbell from?
    Intermediate Answer: New Zealand.
    So the final answer is: No
    """}
]

# 配置一个格式化程序，该格式化程序将prompt格式化为字符串。此格式化程序应该是一个 PromptTemplate 对象。
example_prompt = PromptTemplate(input_variables=["question", "answer"], template="Question: {question}\n{answer}")
print(example_prompt.format(**examples[0]))

# 创建一个选择器来选择最相似的例子
example_selector = SemanticSimilarityExampleSelector(
    examples=examples,
    vector_store=Chroma(),
    embeddings_model=OpenAIEmbeddings(),
    example_prompt=example_prompt
)

# 最后用FewShotPromptTemplate 来创建一个提示词模板，该模板将输入变量作为输入，并将其格式化为包含示例的提示词。
prompt = FewShotPromptTemplate(
    example_selector=example_selector,
    example_prompt=example_prompt,
    suffix="Question: {input}",
    input_variables=["input"]
)
print(prompt)
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.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.

from langchain.prompts import ChatPromptTemplate, FewShotChatMessagePromptTemplate

# 这是一个聊天提示词模板，它将输入变量作为输入，并将其格式化为包含示例的提示词。
examples = [{"input": "2+2", "output": "4"}, {"input": "2+3", "output": "5"},]

# 提示词模板，用于格式化每个单独的示例。
example_prompt = ChatPromptTemplate.from_messages(
    [("human", "{input}"),
     ("ai", "{output}"),])

few_shot_prompt = FewShotChatMessagePromptTemplate(
    example_prompt=example_prompt,
    examples=examples)

print(few_shot_prompt.format())
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.

{
  "_type": "prompt",
  "input_variables": ["adjective", "content"],
  "template": "Tell me a {adjective} joke about {content}."
}
1.
2.
3.
4.
5.

from langchain.prompts import load_prompt

prompt = load_prompt("./simple_prompt.json")
print(prompt.format(adjective="funny", content="chickens"))
1.
2.
3.
4.

{
  "_type": "prompt",
  "input_variables": ["adjective", "content"],
  "template_path": "./simple_template.txt"
}
1.
2.
3.
4.
5.

Tell me a {adjective} joke about {content}.
1.

from langchain.prompts.example_selector import SemanticSimilarityExampleSelector
from langchain.vectorstores import Chroma
from langchain.embeddings import OpenAIEmbeddings

example_selector = SemanticSimilarityExampleSelector.from_examples(
    # 可选的示例列表。
    examples,
    # 用于生成嵌入的嵌入类，这些嵌入用于测量语义相似性。
    OpenAIEmbeddings(),
    # 用于存储嵌入并进行相似性搜索的 VectorStore 类。
    Chroma,
    # 要生成的示例数。
    k=1)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.

# 选择与输入最相似的示例。
question = "Who was the father of Mary Ball Washington?"
selected_examples = example_selector.select_examples({"question": question})
print(f"Examples most similar to the input: {question}")
for example in selected_examples:
    print("\n")
    for k, v in example.items():
        print(f"{k}: {v}")
1.
2.
3.
4.
5.
6.
7.
8.

examples = [
    {"input": "2+2", "output": "4"},
    {"input": "2+3", "output": "5"},
    {"input": "2+4", "output": "6"},
    {"input": "What did the cow say to the moon?", "output": "nothing at all"},
    {
        "input": "Write me a poem about the moon",
        "output": "One for the moon, and one for me, who are we to talk about the moon?",
    },
]

# 由于我们使用向量存储来根据语义相似性选择示例，因此我们需要首先填充存储。
to_vectorize = [" ".join(example.values()) for example in examples]

# 这里就单纯理解为将value对应的值提取出来进行格式化即可。

# 创建向量库后，可以创建 example_selector 以表示返回的相似向量的个数
# 注意：您需要先创建一个向量存储库（例如：vectorstore = ...）并填充它，然后将其传递给 SemanticSimilarityExampleSelector。
example_selector = SemanticSimilarityExampleSelector(vectorstore=vectorstore, k=2)

# 提示词模板将通过将输入传递给 `select_examples` 方法来加载示例
example_selector.select_examples({"input": "horse"})
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.

from langchain.llms import OpenAI
llm = OpenAI()
print(llm('你是谁'))
1.
2.
3.

llm_result = llm.generate(["给我背诵一首古诗", "给我讲个100字小故事"]*10)
1.

import os
import openai
import asyncio
from langchain.llms import OpenAI

# 设置代理
openai.proxy = os.getenv('https_proxy')

# 定义一个同步方式生成文本的函数
def generate_serially():
    llm = OpenAI(temperature=0.9)  # 创建OpenAI对象，并设置temperature参数为0.9
    for _ in range(10):  # 循环10次
        resp = llm.generate(["Hello, how are you?"])  # 调用generate方法生成文本
        print(resp.generations[0][0].text)  # 打印生成的文本

# 定义一个异步生成文本的函数
async def async_generate(llm):
    resp = await llm.agenerate(["Hello, how are you?"])  # 异步调用agenerate方法生成文本
    print(resp.generations[0][0].text)  # 打印生成的文本

# 定义一个并发（异步）方式生成文本的函数
async def generate_concurrently():
    llm = OpenAI(temperature=0.9)  # 创建OpenAI对象，并设置temperature参数为0.9
    tasks = [async_generate(llm) for _ in range(10)]  # 创建10个异步任务
    await asyncio.gather(*tasks)  # 使用asyncio.gather等待所有异步任务完成
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.callbacks.manager import CallbackManagerForLLMRun
from langchain.llms.base import LLM
from typing import Optional, List, Any, Mapping

class CustomLLM(LLM):  # 这个类 CustomLLM 继承了 LLM 类，并增加了一个新的类变量 n。
    n: int  # 类变量，表示一个整数

    @property
    def _llm_type(self) -> str:
        return "custom"

    def _call(
        self,
        prompt: str,  # 输入的提示字符串
        stop: Optional[List[str]] = None,  # 可选的停止字符串列表，默认为 None
        run_manager: Optional[CallbackManagerForLLMRun] = None,  # 可选的回调管理器，默认为 None
        **kwargs: Any,
    ) -> str:
        # 如果 stop 参数不为 None，则抛出 ValueError 异常
        if stop is not None:
            raise ValueError("stop kwargs are not permitted.")
        return prompt[: self.n]  # 返回 prompt 字符串的前 n 个字符

    @property  # 一个属性装饰器，用于获取 _identifying_params 的值
    def _identifying_params(self) -> Mapping[str, Any]:
        """Get the identifying parameters."""  # 这个方法的文档字符串，说明这个方法的功能是获取标识参数
        return {"n": self.n}  # 返回一个字典，包含 n 的值
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.

# 从langchain.llms.fake模块导入FakeListLLM类，此类可能用于模拟或伪造某种行为
from langchain.llms.fake import FakeListLLM
from langchain.agents import load_tools
from langchain.agents import initialize_agent
from langchain.agents import AgentType

# 调用load_tools函数，加载"python_repl"的工具
tools = load_tools(["python_repl"])
# 定义一个响应列表，这些响应可能是模拟LLM的预期响应
responses = ["Action: Python REPL\nAction Input: print(2 + 2)", "Final Answer: 4"]
# 使用上面定义的responses初始化一个FakeListLLM对象
llm = FakeListLLM(responses=responses)
# 调用initialize_agent函数，使用上面的tools和llm，以及指定的代理类型和verbose参数来初始化一个代理
agent = initialize_agent(
    tools, llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True
)
# 调用代理的run方法，传递字符串"whats 2 + 2"作为输入，询问代理2加2的结果
agent.run("whats 2 + 2")
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.

# 从langchain.llms.human模块导入HumanInputLLM类，此类可能允许人类输入或交互来模拟LLM的行为
from langchain.llms.human import HumanInputLLM
from langchain.agents import load_tools
from langchain.agents import initialize_agent
from langchain.agents import AgentType

# 调用load_tools函数，加载名为"wikipedia"的工具
tools = load_tools(["wikipedia"])

# 初始化一个HumanInputLLM对象，其中prompt_func是一个函数，用于打印提示信息
llm = HumanInputLLM(
    prompt_func=lambda prompt: print(f"\n===PROMPT====\n{prompt}\n=====END OF PROMPT======"))
# 调用initialize_agent函数，使用上面的tools和llm，以及指定的代理类型和verbose参数来初始化一个代理
agent = initialize_agent(tools, llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True)
# 调用代理的run方法，传递字符串"What is 'Bocchi the Rock!'?"作为输入，询问代理关于'Bocchi the Rock!'的信息
agent.run("What is 'Bocchi the Rock!'?")
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.

from langchain.cache import SQLiteCache
import langchain
from langchain.llms import OpenAI
import time

langchain.llm_cache = SQLiteCache(database_path=".langchain.db")

llm = OpenAI(model_name="text-davinci-002", n=2, best_of=2)

start_time = time.time()  # 记录开始时间
print(llm.predict("用中文讲个笑话"))
end_time = time.time()  # 记录结束时间
elapsed_time = end_time - start_time  # 计算总时间
print(f"Predict method took {elapsed_time:.4f} seconds to execute.")
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.

from langchain.cache import SQLiteCache
import langchain
from langchain.llms import OpenAI
import time

langchain.llm_cache = SQLiteCache(database_path=".langchain.db")

llm = OpenAI(model_name="text-davinci-002", n=2, best_of=2)

start_time = time.time()  # 记录开始时间
print(llm.predict("用中文讲个笑话"))
end_time = time.time()  # 记录结束时间
elapsed_time = end_time - start_time  # 计算总时间
print(f"Predict method took {elapsed_time:.4f} seconds to execute.")
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.

from langchain.llms import OpenAI
from langchain.callbacks import get_openai_callback

llm = OpenAI(model_name="text-davinci-002", n=2, best_of=2, cache=None)

with get_openai_callback() as cb:
    result = llm("讲个笑话")
    print(cb)
1.
2.
3.
4.
5.
6.
7.
8.

from langchain.agents import load_tools
from langchain.agents import initialize_agent
from langchain.agents import AgentType
from langchain.llms import OpenAI
from langchain.callbacks import get_openai_callback

llm = OpenAI(temperature=0)
tools = load_tools(["llm-math"], llm=llm)
agent = initialize_agent(
    tools, llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True
)

with get_openai_callback() as cb:
    response = agent.run("王菲现在的年龄是多少？")
    print(f"Total Tokens: {cb.total_tokens}")
    print(f"Prompt Tokens: {cb.prompt_tokens}")
    print(f"Completion Tokens: {cb.completion_tokens}")
    print(f"Total Cost (USD): ${cb.total_cost}")
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.

from langchain.llms.loading import load_llm

llm = load_llm("llm.json")

{
  "model_name": "text-davinci-003",
  "temperature": 0.7,
  "max_tokens": 256,
  "top_p": 1.0,
  "frequency_penalty": 0.0,
  "presence_penalty": 0.0,
  "n": 1,
  "best_of": 1,
  "request_timeout": None,
  "_type": "openai"
}
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.

from langchain.llms import OpenAI
from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler

llm = OpenAI(streaming=True, callbacks=[StreamingStdOutCallbackHandler()], temperature=0)
resp = llm("Write me a song about sparkling water.")
1.
2.
3.
4.
5.

from langchain.output_parsers import CommaSeparatedListOutputParser
from langchain.prompts import PromptTemplate, ChatPromptTemplate, HumanMessagePromptTemplate
from langchain.llms import OpenAI
from langchain.chat_models 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 = OpenAI(temperature=0)

_input = prompt.format(subject="冰淇淋口味")
output = model(_input)

output_parser.parse(output)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.

from langchain.prompts import PromptTemplate
from langchain.output_parsers import DatetimeOutputParser
from langchain.chains import LLMChain
from langchain.llms import OpenAI

output_parser = DatetimeOutputParser()

template = """回答用户的问题:
{question}
{format_instructions}"""

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

chain = LLMChain(prompt=prompt, llm=OpenAI())

output = chain.run("bitcoin是什么时候成立的？用英文格式输出时间")
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.

from langchain.output_parsers.enum import EnumOutputParser
from enum import Enum

class Colors(Enum):
    RED = "red"
    GREEN = "green"
    BLUE = "blue"

parser = EnumOutputParser(enum=Colors)
1.
2.
3.
4.
5.
6.
7.
8.
9.

# 导入所需的库和模块
from langchain.prompts import PromptTemplate, ChatPromptTemplate, HumanMessagePromptTemplate
from langchain.llms import OpenAI
from langchain.chat_models import ChatOpenAI
from langchain.output_parsers import PydanticOutputParser
from pydantic import BaseModel, Field, validator
from typing import List

# 定义一个表示演员的数据结构，包括他们的名字和他们出演的电影列表
class Actor(BaseModel):
    name: str = Field(description="name of an actor")  # 演员的名字
    film_names: List[str] = Field(description="list of names of films they starred in")  # 他们出演的电影列表

# 定义一个查询，用于提示生成随机演员的电影作品列表
actor_query = "Generate the filmography for a random actor."

# 使用`Actor`模型初始化解析器
parser = PydanticOutputParser(pydantic_object=Actor)

# 定义一个格式错误的字符串数据
misformatted = "{'name': 'Tom Hanks', 'film_names': ['Forrest Gump']}"

# 使用解析器尝试解析上述数据
try:
    parsed_data = parser.parse(misformatted)
except Exception as e:
    print(f"Error: {e}")
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.

from langchain.output_parsers import RetryWithErrorOutputParser
from langchain.llms import OpenAI

retry_parser = RetryWithErrorOutputParser.from_llm(
    parser=parser, llm=OpenAI(temperature=0))
retry_parser.parse_with_prompt(bad_response, prompt_value)
1.
2.
3.
4.
5.
6.

from langchain.document_loaders import TextLoader
# 创建一个TextLoader实例，指定要加载的Markdown文件路径
loader = TextLoader("./index.md")
# 使用load方法加载文件内容并打印
print(loader.load())
1.
2.
3.
4.
5.

# 导入CSVLoader类
from langchain.document_loaders.csv_loader import CSVLoader

# 创建CSVLoader实例，指定要加载的CSV文件路径
loader = CSVLoader(file_path='./index.csv')

# 使用load方法加载数据并将其存储在数据变量中
data = loader.load()
1.
2.
3.
4.
5.
6.
7.
8.

from langchain.document_loaders.csv_loader import CSVLoader

# 创建CSVLoader实例，指定要加载的CSV文件路径和CSV参数
loader = CSVLoader(file_path='./index.csv', csv_args={
    'delimiter': ',',
    'quotechar': '"',
    'fieldnames': ['title', 'content']
})

# 使用load方法加载数据并将其存储在数据变量中
data = loader.load()
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.

from langchain.document_loaders.csv_loader import CSVLoader

# 创建CSVLoader实例，指定要加载的CSV文件路径和源列名
loader = CSVLoader(file_path='./index.csv', source_column="context")

# 使用load方法加载数据并将其存储在数据变量中
data = loader.load()
1.
2.
3.
4.
5.
6.
7.

# 导入DirectoryLoader类
from langchain.document_loaders import DirectoryLoader

# 创建DirectoryLoader实例，指定要加载的文件夹路径、要加载的文件类型和是否使用多线程
loader = DirectoryLoader('/Users/kyoku/Desktop/LLM/documentstore', glob='**/*.md', use_multithreading=True)

# 使用load方法加载所有文档并将其存储在docs变量中
docs = loader.load()

# 打印加载的文档数量
print(len(docs))

# 导入UnstructuredHTMLLoader类
from langchain.document_loaders import UnstructuredHTMLLoader

# 创建UnstructuredHTMLLoader实例，指定要加载的HTML文件路径
loader = UnstructuredHTMLLoader("./index.html")

# 使用load方法加载HTML文件内容并将其存储在data变量中
data = loader.load()

# 导入BSHTMLLoader类
from langchain.document_loaders import BSHTMLLoader

# 创建BSHTMLLoader实例，指定要加载的HTML文件路径
loader = BSHTMLLoader("./index.html")

# 使用load方法加载HTML文件内容并将其存储在data变量中
data = loader.load()
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.

# 打开一个文本文件并读取内容
with open('./test.txt') as f:
    state_of_the_union = f.read()

# 导入RecursiveCharacterTextSplitter类
from langchain.text_splitter import RecursiveCharacterTextSplitter

# 创建RecursiveCharacterTextSplitter实例，设置块大小、块重叠、长度函数和是否添加开始索引
text_splitter = RecursiveCharacterTextSplitter(
    chunk_size=100,
    chunk_overlap=20,
    length_function=len,
    add_start_index=True,
)

# 使用create_documents方法创建文档并将其存储在texts变量中
texts = text_splitter.create_documents([state_of_the_union])
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.

# 导入所需的类和枚举
from langchain.text_splitter import RecursiveCharacterTextSplitter, Language

# 定义一个包含Python代码的字符串
PYTHON_CODE = """
def hello_world():
    print("Hello, World!")

# Call the function
hello_world()
"""

# 使用from_language方法创建一个针对Python语言的RecursiveCharacterTextSplitter实例
python_splitter = RecursiveCharacterTextSplitter.from_language(
    language=Language.PYTHON, chunk_size=50, chunk_overlap=0
)

# 使用create_documents方法创建文档并将其存储在python_docs变量中
python_docs = python_splitter.create_documents([PYTHON_CODE])
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.

text = """客户：您好，我想咨询一下信用卡的问题。\n客服：您好，欢迎咨询建行信用卡，我是客服小李，请问有什么问题我可以帮您解答吗？\n客户：我想了解一下信用卡的年费如何收取？\n客服：关于信用卡年费的收取，我们会在每年的固定日期为您的信用卡收取年费。当然，如果您在一年内的消费达到一定金额，年费会自动免除。具体的免年费标准，请您查看信用卡合同条款或登录我们的网站查询。\n客户：好的，谢谢。那我还想问一下，如何提高信用卡的额度？\n客服：关于提高信用卡额度，您可以通过以下途径操作：1. 登录建行信用卡官方网站或手机APP，提交在线提额申请；2. 拨打我们的客服热线，按语音提示进行提额申请；3. 您还可以前往附近的建行网点，提交提额申请。在您提交申请后，我们会根据您的信用状况进行审核，审核通过后，您的信用卡额度将会相应提高。\n客户：明白了，非常感谢您的解答。\n客服：您太客气了，很高兴能够帮到您。如果您还有其他问题，请随时联系我们。祝您生活愉快！"""
list_text = text.split('\n')

from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.vectorstores import FAISS

db = FAISS.from_texts(list_text, OpenAIEmbeddings())

query = "信用卡的额度可以提高吗"
docs = db.similarity_search(query)
print(docs[0].page_content)

embedding_vector = OpenAIEmbeddings().embed_query(query)
print(f'embedding_vector：{embedding_vector}')
docs = db.similarity_search_by_vector(embedding_vector)
print(docs[0].page_content)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.

# 使用带分数的相似性搜索
docs_and_scores = db.similarity_search_with_score(query)

# 打印文档及其相似性分数
for doc, score in docs_and_scores:
    print(f"Document: {doc.page_content}\nScore: {score}\n")
1.
2.
3.
4.
5.
6.

# 保存
db.save_local("faiss_index")
# 加载
new_db = FAISS.load_local("faiss_index", OpenAIEmbeddings())
1.
2.
3.
4.

memory = ConversationBufferMemory() memory.save_context({"input": "你好，我是人类"}, {"output": "你好，我是AI助手"})memory.save_context({"input": "很开心认识你"}, {"output": "我也是"})
1.

print(memory.load_memory_variables({}))
# {'history': 'Human: 你好，我是人类\nAI: 你好，我是AI助手\nHuman: 很开心认识你\nAI: 我也是'}
1.
2.

memory = ConversationBufferWindowMemory(k=1)
memory.save_context({"input": "你好，我是人类"}, {"output": "你好，我是AI助手"})
memory.save_context({"input": "很开心认识你"}, {"output": "我也是"})
1.
2.
3.

print(memory.load_memory_variables({}))
# {'history': 'Human: 很开心认识你\nAI: 我也是'}
1.
2.

from langchain.chat_models import ChatOpenAI
from langchain.memory import ConversationTokenBufferMemory
llm = ChatOpenAI(temperature=0.0)
memory = ConversationTokenBufferMemory(llm=llm,)
memory.save_context({"input": "春眠不觉晓"}, {"output": "处处闻啼鸟"})
memory.save_context({"input": "夜来风雨声"}, {"output": "花落知多少"})
print(memory.load_memory_variables({}))
#{'history': 'AI: 花落知多少。'}
1.
2.
3.
4.
5.
6.
7.
8.

memory = ConversationSummaryBufferMemory(llm=llm, max_token_limit=40, return_messages=True)
memory.save_context({"input": "嗨"}, {"output": "你好吗"})
memory.save_context({"input": "没什么特别的，你呢"}, {"output": "我也是"})

messages = memory.chat_memory.messages
previous_summary = ""
print(memory.predict_new_summary(messages, previous_summary))
# 人类和AI都表示没有做什么特别的事
1.
2.
3.
4.
5.
6.
7.
8.

import faiss

from langchain.docstore import InMemoryDocstore
from langchain.vectorstores import FAISS

embedding_size = 1536 # Dimensions of the OpenAIEmbeddings
index = faiss.IndexFlatL2(embedding_size)
embedding_fn = OpenAIEmbeddings().embed_query
vectorstore = FAISS(embedding_fn, index, InMemoryDocstore({}), {})
1.
2.
3.
4.
5.
6.
7.
8.
9.

# 在实际使用中，可以将`k` 设为更高的值，这里使用 k=1 来展示
# 向量查找仍然返回语义相关的信息
retriever = vectorstore.as_retriever(search_kwargs=dict(k=1))
memory = VectorStoreRetrieverMemory(retriever=retriever)

# 当添加到一个代理时，内存对象可以保存来自对话或使用的工具的相关信息
memory.save_context({"input": "我最喜欢的食物是披萨"}, {"output": "好的，我知道了"})
memory.save_context({"input": "我最喜欢的运动是足球"}, {"output": "..."})
memory.save_context({"input": "我不喜欢凯尔特人队"}, {"output": "好的"}) 
print(memory.load_memory_variables({"prompt": "我应该看什么运动?"})["history"])
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.

{
  'history': [
    {
      'input': '我最喜欢的运动是足球',
      'output': '...'
    }
  ]
}
1.
2.
3.
4.
5.
6.
7.
8.

llm = OpenAI(temperature=0) # 可以是任何有效的LLM
_DEFAULT_TEMPLATE = """以下是一个人类与AI之间的友好对话。AI非常健谈，并从其上下文中提供大量具体细节。如果AI不知道问题的答案，它会诚实地说不知道。

之前对话的相关部分：
{history}

（如果不相关，您不需要使用这些信息）

当前对话：
人类：{input}
AI："""
PROMPT = PromptTemplate(
    input_variables=["history", "input"], template=_DEFAULT_TEMPLATE
)
conversation_with_summary = ConversationChain(
    llm=llm, 
    prompt=PROMPT,
    # 我们为测试目的设置了一个非常低的max_token_limit。
    memory=memory,
    verbose=True
)
conversation_with_summary.predict(input="嗨，我叫Perry，你好吗？")
# 输出："> Entering new ConversationChain chain...
# Prompt after formatting:
# ...
# > Finished chain.
# " 嗨，Perry，我很好。你呢？"

# 这里，与篮球相关的内容被提及
conversation_with_summary.predict(input="我最喜欢的运动是什么？")
# 输出："> Entering new ConversationChain chain...
# ...
# > Finished chain.
# ' 你之前告诉我你最喜欢的运动是足球。'"

# 尽管语言模型是无状态的，但由于获取到了相关的记忆，它可以“推理”出时间。
# 为记忆和数据加上时间戳通常是有用的，以便让代理确定时间相关性
conversation_with_summary.predict(input="我的最喜欢的食物是什么？")
# 输出："> Entering new ConversationChain chain...
# ...
# > Finished chain.
# ' 你说你最喜欢的食物是披萨。'"

# 对话中的记忆被自动存储，
# 由于这个查询与上面的介绍聊天最匹配，
# 代理能够“记住”用户的名字。
conversation_with_summary.predict(input="我的名字是什么？")
# 输出："> Entering new ConversationChain chain...
# ...
# > Finished chain.
# ' 你的名字是Perry。'"
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.
48.
49.
50.
51.

from langchain import PromptTemplate, OpenAI, LLMChain

prompt_template = "What is a good name for a company that makes {product}?"

llm = OpenAI(temperature=0)
chain = LLMChain(
    llm=llm,
    prompt=PromptTemplate.from_template(prompt_template)
)
print(chain("colorful socks")) 
# 输出结果'Socktastic!'
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.

# This is an LLMChain to write a synopsis given a title of a play.
from langchain import PromptTemplate, OpenAI, LLMChain

llm = OpenAI(temperature=.7)
template = """You are a playwright. Given the title of play, it is your job to write a synopsis for that title.

Title: {title}
Playwright: This is a synopsis for the above play:"""
prompt_template = PromptTemplate(input_variables=["title"], template=template)
synopsis_chain = LLMChain(llm=llm, prompt=prompt_template)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.

# This is an LLMChain to write a review of a play given a synopsis.
from langchain import PromptTemplate, OpenAI, LLMChain

llm = OpenAI(temperature=.7)
template = """You are a play critic from the New York Times. Given the synopsis of play, it is your job to write a review for that play.

Play Synopsis:
{synopsis}
Review from a New York Times play critic of the above play:"""
prompt_template = PromptTemplate(input_variables=["synopsis"], template=template)
review_chain = LLMChain(llm=llm, prompt=prompt_template)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.

from langchain.chains import SimpleSequentialChain
overall_chain = SimpleSequentialChain(chains=[synopsis_chain, review_chain], verbose=True)
print(review = overall_chain.run("Tragedy at sunset on the beach"))
1.
2.
3.

from langchain import PromptTemplate, OpenAI, LLMChain

llm = OpenAI(temperature=.7)
template = """You are a playwright. Given the title of play and the era it is set in, it is your job to write a synopsis for that title.

Title: {title}
Era: {era}
Playwright: This is a synopsis for the above play:"""
prompt_template = PromptTemplate(input_variables=["title", 'era'], template=template)
synopsis_chain = LLMChain(llm=llm, prompt=prompt_template, output_key="synopsis")
#第一条chain
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.

from langchain import PromptTemplate, OpenAI, LLMChain

llm = OpenAI(temperature=.7)
template = """You are a play critic from the New York Times. Given the synopsis of play, it is your job to write a review for that play.

Play Synopsis:
{synopsis}
Review from a New York Times play critic of the above play:"""
prompt_template = PromptTemplate(input_variables=["synopsis"], template=template)
review_chain = LLMChain(llm=llm, prompt=prompt_template, output_key="review")
#第二条chain
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.

from langchain.chains import SequentialChain

overall_chain = SequentialChain(
    chains=[synopsis_chain, review_chain],
    input_variables=["era", "title"],
    # Here we return multiple variables
    output_variables=["synopsis", "review"],
    verbose=True)
#第三条chain
1.
2.
3.
4.
5.
6.
7.
8.
9.

overall_chain({"title": "Tragedy at sunset on the beach", "era": "Victorian England"})
1.

physics_template = """You are a very smart physics professor. \
You are great at answering questions about physics in a concise and easy to understand manner. \
When you don't know the answer to a question you admit that you don't know.

Here is a question:
{input}"""

math_template = """You are a very good mathematician. You are great at answering math questions. \
You are so good because you are able to break down hard problems into their component parts, \
answer the component parts, and then put them together to answer the broader question.

Here is a question:
{input}"""
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.

prompt_infos = [
    {
        "name": "physics",
        "description": "Good for answering questions about physics",
        "prompt_template": physics_template
    },
    {
        "name": "math",
        "description": "Good for answering math questions",
        "prompt_template": math_template
    }
]
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.

from langchain import ConversationChain, LLMChain, PromptTemplate, OpenAI
llm = OpenAI()
destination_chains = {}
for p_info in prompt_infos:
    name = p_info["name"]
    prompt_template = p_info["prompt_template"]
    prompt = PromptTemplate(template=prompt_template, input_variables=["input"])
    chain = LLMChain(llm=llm, prompt=prompt)
    destination_chains[name] = chain

default_chain = ConversationChain(llm=llm, output_key="text")
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.

from langchain.chains.router.llm_router import LLMRouterChain, RouterOutputParser
from langchain.chains.router.multi_prompt_prompt import MULTI_PROMPT_ROUTER_TEMPLATE

# Create a list of destinations
destinations = [f"{p['name']}: {p['description']}" for p in prompt_infos]
destinations_str = "\n".join(destinations)

# Create a router template
router_template = MULTI_PROMPT_ROUTER_TEMPLATE.format(destinations=destinations_str)


router_prompt = PromptTemplate(
    template=router_template,
    input_variables=["input"],
    output_parser=RouterOutputParser(),
)

router_chain = LLMRouterChain.from_llm(llm, router_prompt)
chain = MultiPromptChain(
    router_chain=router_chain,
    destination_chains=destination_chains,
    default_chain=default_chain,
    verbose=True,
)
print(chain.run('什么是黑体辐射'))
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.

# 初始化 agent
agent = initialize_agent(
    tools,  # 配置工具集
    llm,  # 配置大语言模型 负责决策
    agent=AgentType.OPENAI_FUNCTIONS,  # 设置 agent 类型 
    agent_kwargs=agent_kwargs,  # 设定 agent 角色
    verbose=True,
    memory=memory, # 配置记忆模式 )
1.
2.
3.
4.
5.
6.
7.
8.

from langchain.agents import initialize_agent, Tool
tools = [
    Tool(
        name="Search",
        func=search,
        description="useful for when you need to answer questions about current events, data. You should ask targeted questions"
    ),
    ScrapeWebsiteTool(),
]
1.
2.
3.
4.
5.
6.
7.
8.
9.

# 调用 Google search by Serper
def search(query):
    serper_google_url = os.getenv("SERPER_GOOGLE_URL")

    payload = json.dumps({
        "q": query
    })

    headers = {
        'X-API-KEY': serper_api_key,
        'Content-Type': 'application/json'
    }

    response = requests.request("POST", serper_google_url, headers=headers, data=payload)

    print(f'Google 搜索结果: \n {response.text}')
    return response.text
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.

class ScrapeWebsiteTool(BaseTool):
    name = "scrape_website"
    description = "useful when you need to get data from a website url, passing both url and objective to the function; DO NOT make up any url, the url should only be from the search results"
    args_schema: Type[BaseModel] = ScrapeWebsiteInput

    def _run(self, target: str, url: str):
        return scrape_website(target, url)

    def _arun(self, url: str):
        raise NotImplementedError("error here")
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.

class ScrapeWebsiteInput(BaseModel):
    """Inputs for scrape_website"""
    target: str = Field(
        description="The objective & task that users give to the agent")
    url: str = Field(description="The url of the website to be scraped")
1.
2.
3.
4.
5.

# 根据 url 爬取网页内容，给出最终解答
# target ：分配给 agent 的初始任务
# url ： Agent 在完成以上目标时所需要的URL，完全由Agent自主决定并且选取，其内容或是中间步骤需要，或是最终解答需要
def scrape_website(target: str, url: str):
    print(f"开始爬取： {url}...")

    headers = {
        'Cache-Control': 'no-cache',
        'Content-Type': 'application/json',
    }

    payload = json.dumps({
        "url": url
    })

    post_url = f"https://chrome.browserless.io/content?token={browserless_api_key}"
    response = requests.post(post_url, headers=headers, data=payload)

    # 如果返回成功
    if response.status_code == 200:
        soup = BeautifulSoup(response.content, "html.parser")
        text = soup.get_text()
        print("爬取的具体内容:", text)

        # 控制返回内容长度，如果内容太长就需要切片分别总结处理
        if len(text) > 5000:
            # 总结爬取的返回内容
            output = summary(target, text)
            return output
        else:
            return text
    else:
        print(f"HTTP请求错误，错误码为{response.status_code}")
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.

# 如果需要处理的内容过长，先切片分别处理，再综合总结
# 使用 Map-Reduce 方式
def summary(target, content):
    # model list ： https://platform.openai.com/docs/models
    # gpt-4-32k   gpt-3.5-turbo-16k-0613
    llm = ChatOpenAI(temperature=0, model="gpt-3.5-turbo-16k-0613")

    # 定义大文本切割器
    # chunk_overlap 是一个在使用 OpenAI 的 GPT-3 或 GPT-4 API 时可能会遇到的参数，特别是需要处理长文本时。
    # 该参数用于控制文本块（chunks）之间的重叠量。
    # 上下文维护：重叠确保模型在处理后续块时有足够的上下文信息。
    # 连贯性：它有助于生成更连贯和一致的输出，因为模型可以“记住”前一个块的部分内容。
    text_splitter = RecursiveCharacterTextSplitter(
        separators=["\n\n", "\n"], chunk_size=5000, chunk_overlap=200)

    docs = text_splitter.create_documents([content])
    map_prompt = """
    Write a summary of the following text for {target}:
    "{text}"
    SUMMARY:
    """
    map_prompt_template = PromptTemplate(
        template=map_prompt, input_variables=["text", "target"])

    summary_chain
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.

# 初始化大语言模型，负责决策
llm = ChatOpenAI(temperature=0, model="gpt-3.5-turbo-16k-0613")
1.
2.

# 初始化agents的详细描述
system_message = SystemMessage(
    content="""您是一位世界级的研究员，可以对任何主题进行详细研究并产生基于事实的结果；
            您不会凭空捏造事实，您会尽最大努力收集事实和数据来支持研究。

            请确保按照以下规则完成上述目标：
            1/ 您应该进行足够的研究，尽可能收集关于目标的尽可能多的信息
            2/ 如果有相关链接和文章的网址，您将抓取它以收集更多信息
            3/ 在抓取和搜索之后，您应该思考“根据我收集到的数据，是否有新的东西需要我搜索和抓取以提高研究质量？”如果答案是肯定的，继续；但不要进行超过5次迭代
            4/ 您不应该捏造事实，您只应该编写您收集到的事实和数据
            5/ 在最终输出中，您应该包括所有参考数据和链接以支持您的研究；您应该包括所有参考数据和链接以支持您的研究
            6/ 在最终输出中，您应该包括所有参考数据和链接以支持您的研究；您应该包括所有参考数据和链接以支持您的研究"""
)
# 初始化 agent 角色模板
agent_kwargs = {
    "extra_prompt_messages": [MessagesPlaceholder(variable_name="memory")],
    "system_message": system_message,
}

# 初始化记忆类型
memory = ConversationSummaryBufferMemory(
    memory_key="memory", return_messages=True, llm=llm, max_token_limit=300)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.

# 初始化记忆类型
memory = ConversationSummaryBufferMemory(
    memory_key="memory", return_messages=True, llm=llm, max_token_limit=300)
1.
2.
3.

import os
from dotenv import load_dotenv

from langchain import PromptTemplate
from langchain.agents import initialize_agent, Tool
from langchain.agents import AgentType
from langchain.chat_models import ChatOpenAI
from langchain.prompts import MessagesPlaceholder
from langchain.memory import ConversationSummaryBufferMemory
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.chains.summarize import load_summarize_chain
from langchain.tools import BaseTool
from pydantic import BaseModel, Field
from langchain.schema import SystemMessage

from typing import Type
from bs4 import BeautifulSoup
import requests
import json

import streamlit as st

# 加载必要的参数
load_dotenv()
serper_api_key=os.getenv("SERPER_API_KEY")
browserless_api_key=os.getenv("BROWSERLESS_API_KEY")
openai_api_key=os.getenv("OPENAI_API_KEY")
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.

def main():
    st.set_page_config(page_title="AI Assistant Agent", page_icon=":dolphin:")

    st.header("LangChain 实例讲解 3 -- Agent", divider='rainbow')
    st.header("AI Agent :blue[助理] :dolphin:")

    query = st.text_input("请提问题和需求：")

    if query:
        st.write(f"开始收集和总结资料 【 {query}】 请稍等")

        result = agent({"input": query})

        st.info(result['output'])
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.

class BaseCallbackHandler:
    """Base callback handler that can be used to handle callbacks from langchain."""

    def on_llm_start(
        self, serialized: Dict[str, Any], prompts: List[str], **kwargs: Any
    ) -> Any:
        """Run when LLM starts running."""

    def on_chat_model_start(
        self, serialized: Dict[str, Any], messages: List[List[BaseMessage]], **kwargs: Any
    ) -> Any:
        """Run when Chat Model starts running."""

    def on_llm_new_token(self, token: str, **kwargs: Any) -> Any:
        """Run on new LLM token. Only available when streaming is enabled."""

    def on_llm_end(self, response: LLMResult, **kwargs: Any) -> Any:
        """Run when LLM ends running."""

    def on_llm_error(
        self, error: Union[Exception, KeyboardInterrupt], **kwargs: Any
    ) -> Any:
        """Run when LLM errors."""

    def on_chain_start(
        self, serialized: Dict[str, Any], inputs: Dict[str, Any], **kwargs: Any
    ) -> Any:
        """Run when chain starts running."""

    def on_chain_end(self, outputs: Dict[str, Any], **kwargs: Any) -> Any:
        """Run when chain ends running."""

    def on_chain_error(
        self, error: Union[Exception, KeyboardInterrupt], **kwargs: Any
    ) -> Any:
        """Run when chain errors."""

    def on_tool_start(
        self, serialized: Dict[str, Any], input_str: str, **kwargs: Any
    ) -> Any:
        """Run when tool starts running."""

    def on_tool_end(self, output: str, **kwargs: Any) -> Any:
        """Run when tool ends running."""

    def on_tool_error(
        self, error: Union[Exception, KeyboardInterrupt], **kwargs: Any
    ) -> Any:
        """Run when tool errors."""

    def on_text(self, text: str, **kwargs: Any) -> Any:
        """Run on arbitrary text."""

    def on_agent_action(self, action: AgentAction, **kwargs: Any) -> Any:
        """Run on agent action."""

    def on_agent_finish(self, finish: AgentFinish, **kwargs: Any) -> Any:
        """Run on agent end."""
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.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.

from langchain.callbacks import StdOutCallbackHandler
from langchain.chains import LLMChain
from langchain.llms import OpenAI
from langchain.prompts import PromptTemplate

handler = StdOutCallbackHandler()
llm = OpenAI()
prompt = PromptTemplate.from_template("Who is {name}?")
chain = LLMChain(llm=llm, prompt=prompt, callbacks=[handler])
chain.run(name="Super Mario")
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.

from langchain.callbacks.base import BaseCallbackHandler
import time

class TimerHandler(BaseCallbackHandler):

    def __init__(self) -> None:
        super().__init__()
        self.previous_ms = None
        self.durations = []

    def current_ms(self):
        return int(time.time() * 1000 + time.perf_counter() % 1 * 1000)

    def on_chain_start(self, serialized, inputs, **kwargs) -> None:
        self.previous_ms = self.current_ms()

    def on_chain_end(self, outputs, **kwargs) -> None:
        if self.previous_ms:
          duration = self.current_ms() - self.previous_ms
          self.durations.append(duration)

    def on_llm_start(self, serialized, prompts, **kwargs) -> None:
        self.previous_ms = self.current_ms()

    def on_llm_end(self, response, **kwargs) -> None:
        if self.previous_ms:
          duration = self.current_ms() - self.previous_ms
          self.durations.append(duration)

llm = OpenAI()
timerHandler = TimerHandler()
prompt = PromptTemplate.from_template("What is the HEX code of color {color_name}?")
chain = LLMChain(llm=llm, prompt=prompt, callbacks=[timerHandler])
response = chain.run(color_name="blue")
print(response)
response = chain.run(color_name="purple")
print(response)
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.

timerHandler = TimerHandler()
llm = OpenAI(callbacks=[timerHandler]) 
response = llm.predict("What is the HEX code of color BLACK?") print(response) 
1.
2.
3.

timerHandler = TimerHandler()
llm = OpenAI()
response = llm.predict("What is the HEX code of color BLACK?", callbacks=[timerHandler])
print(response)
1.
2.
3.
4.