聊天引擎 - 浓缩上下文模式¶

这是一个基于数据检索器的多步骤聊天模式。

对于每次聊天交互：

• 首先将对话和最新用户消息浓缩为一个独立问题
• 然后通过检索器为该独立问题构建上下文
• 最后将上下文连同提示和用户消息传递给大语言模型（LLM）以生成响应

这种方法简单直接，适用于与知识库直接相关的问题及常规交互场景。

如果您在 Colab 上打开此 Notebook，可能需要安装 LlamaIndex 🦙。

In [ ]:

%pip install llama-index-llms-openai

%pip install llama-index-llms-openai

In [ ]:

!pip install llama-index

!pip install llama-index

下载数据¶

In [ ]:

!mkdir -p 'data/paul_graham/'
!wget 'https://raw.githubusercontent.com/run-llama/llama_index/main/docs/docs/examples/data/paul_graham/paul_graham_essay.txt' -O 'data/paul_graham/paul_graham_essay.txt'

!mkdir -p 'data/paul_graham/' !wget 'https://raw.githubusercontent.com/run-llama/llama_index/main/docs/docs/examples/data/paul_graham/paul_graham_essay.txt' -O 'data/paul_graham/paul_graham_essay.txt'

5 行代码快速入门¶

加载数据并构建索引

In [ ]:

import openai
import os

os.environ["OPENAI_API_KEY"] = "sk-..."
openai.api_key = os.environ["OPENAI_API_KEY"]

import openai import os os.environ["OPENAI_API_KEY"] = "sk-..." openai.api_key = os.environ["OPENAI_API_KEY"]

In [ ]:

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader
from llama_index.llms.openai import OpenAI


llm = OpenAI(model="gpt-3.5-turbo")
data = SimpleDirectoryReader(input_dir="./data/paul_graham/").load_data()
index = VectorStoreIndex.from_documents(data)

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader from llama_index.llms.openai import OpenAI llm = OpenAI(model="gpt-3.5-turbo") data = SimpleDirectoryReader(input_dir="./data/paul_graham/").load_data() index = VectorStoreIndex.from_documents(data)

配置聊天引擎

由于检索到的上下文可能会占用大量可用的大语言模型（LLM）上下文空间，我们需要确保为聊天历史记录设置一个较小的容量限制！

In [ ]:

from llama_index.core.memory import ChatMemoryBuffer

memory = ChatMemoryBuffer.from_defaults(token_limit=3900)

chat_engine = index.as_chat_engine(
    chat_mode="condense_plus_context",
    memory=memory,
    llm=llm,
    context_prompt=(
        "You are a chatbot, able to have normal interactions, as well as talk"
        " about an essay discussing Paul Grahams life."
        "Here are the relevant documents for the context:\n"
        "{context_str}"
        "\nInstruction: Use the previous chat history, or the context above, to interact and help the user."
    ),
    verbose=False,
)

from llama_index.core.memory import ChatMemoryBuffer memory = ChatMemoryBuffer.from_defaults(token_limit=3900) chat_engine = index.as_chat_engine( chat_mode="condense_plus_context", memory=memory, llm=llm, context_prompt=( "You are a chatbot, able to have normal interactions, as well as talk" " about an essay discussing Paul Grahams life." "Here are the relevant documents for the context:\n" "{context_str}" "\nInstruction: Use the previous chat history, or the context above, to interact and help the user." ), verbose=False, )

与你的数据对话

In [ ]:

response = chat_engine.chat("What did Paul Graham do growing up")

response = chat_engine.chat("What did Paul Graham do growing up")

In [ ]:

print(response)

print(response)

Growing up, Paul Graham had two main interests: writing and programming. He started by writing short stories, although he admits that they were not very good. In terms of programming, he began working with computers in 9th grade when he had access to an IBM 1401 at his school. He used an early version of Fortran to write programs on punch cards for the 1401. However, he found it challenging to figure out what to do with the machine since the only input option was data stored on punched cards. Later, with the advent of microcomputers, he became more involved in programming and got his own TRS-80 computer. He wrote simple games, a program to predict rocket heights, and even a word processor. Despite his interest in programming, he initially planned to study philosophy in college but eventually switched to AI.

提出后续问题

In [ ]:

response_2 = chat_engine.chat("Can you tell me more?")

response_2 = chat_engine.chat("Can you tell me more?")

In [ ]:

print(response_2)

print(response_2)

Certainly! After Paul Graham switched his focus from philosophy to AI in college, he became fascinated with the field. AI, or artificial intelligence, was gaining attention in the mid-1980s, and Graham was particularly inspired by a novel called "The Moon is a Harsh Mistress" by Robert A. Heinlein, which featured an intelligent computer named Mike. He also saw a PBS documentary showcasing Terry Winograd using a program called SHRDLU, which further fueled his interest in AI.

Graham believed that AI was on the verge of significant advancements, and he wanted to be a part of it. He saw the potential for intelligent machines and the impact they could have on society. This passion for AI led him to pursue it as a field of study and work.

Throughout his journey, Graham continued to explore programming and writing. He wrote numerous essays on various topics and even published a collection of them in a book called "Hackers & Painters." He also worked on spam filters and continued his interest in painting.

Graham's experiences with writing, programming, and his fascination with AI eventually led him to co-found Viaweb, an early e-commerce platform, which was later acquired by Yahoo. This venture marked the beginning of his entrepreneurial career and set the stage for his future endeavors, including the founding of Y Combinator, a renowned startup accelerator.

Overall, Graham's upbringing and early interests in writing and programming played a significant role in shaping his path, leading him to become a successful entrepreneur, investor, and influential figure in the tech industry.

重置会话状态

In [ ]:

chat_engine.reset()

chat_engine.reset()

In [ ]:

response = chat_engine.chat("Hello! What do you know?")

response = chat_engine.chat("Hello! What do you know?")

In [ ]:

print(response)

print(response)

Hello! I'm a chatbot here to help you with any questions or topics you'd like to discuss. Is there something specific you'd like to know or talk about?

流式传输支持¶

In [ ]:

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader
from llama_index.llms.openai import OpenAI
from llama_index.core import Settings

llm = OpenAI(model="gpt-3.5-turbo", temperature=0)
Settings.llm = llm
data = SimpleDirectoryReader(input_dir="./data/paul_graham/").load_data()
index = VectorStoreIndex.from_documents(data)

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader from llama_index.llms.openai import OpenAI from llama_index.core import Settings llm = OpenAI(model="gpt-3.5-turbo", temperature=0) Settings.llm = llm data = SimpleDirectoryReader(input_dir="./data/paul_graham/").load_data() index = VectorStoreIndex.from_documents(data)

In [ ]:

chat_engine = index.as_chat_engine(
    chat_mode="condense_plus_context",
    context_prompt=(
        "You are a chatbot, able to have normal interactions, as well as talk"
        " about an essay discussing Paul Grahams life."
        "Here are the relevant documents for the context:\n"
        "{context_str}"
        "\nInstruction: Based on the above documents, provide a detailed answer for the user question below."
    ),
)

chat_engine = index.as_chat_engine( chat_mode="condense_plus_context", context_prompt=( "You are a chatbot, able to have normal interactions, as well as talk" " about an essay discussing Paul Grahams life." "Here are the relevant documents for the context:\n" "{context_str}" "\nInstruction: Based on the above documents, provide a detailed answer for the user question below." ), )

In [ ]:

response = chat_engine.stream_chat("What did Paul Graham do after YC?")
for token in response.response_gen:
    print(token, end="")

response = chat_engine.stream_chat("What did Paul Graham do after YC?") for token in response.response_gen: print(token, end="")

After Y Combinator (YC), Paul Graham made a significant decision to step back from his role in YC and pursue other endeavors. In 2012, his mother had a stroke caused by colon cancer, which led him to reevaluate his priorities. He realized that YC was consuming more of his attention and that he was ready to hand over the reins to someone else.

Paul approached Jessica Livingston, his wife and co-founder of YC, to see if she wanted to become the president of YC, but she declined. They then decided to recruit Sam Altman, who was a successful entrepreneur and had been involved with YC as a founder. They also involved Robert Morris and Trevor Blackwell, who were original partners in YC.

To ensure the long-term success of YC, they decided to reorganize the company. Previously, YC had been controlled by the original LLC formed by the four founders. However, they wanted to ensure that YC would continue to thrive even without their direct control. If Sam accepted the offer, he would be given the opportunity to reorganize YC, with Paul and Robert retiring and Jessica and Trevor becoming ordinary partners.

After persistent persuasion, Sam agreed to become the president of YC in October 2013, and the transition began with the winter 2014 batch. During the rest of 2013, Paul gradually handed over the responsibilities of running YC to Sam, allowing him to learn the job while Paul focused on his mother, who was battling cancer.

It is worth noting that Paul Graham's decision to step back from YC was influenced by the advice of Robert Morris, who suggested that YC should not be the last cool thing Paul does. This advice made Paul realize that he wanted to explore other opportunities and not be solely defined by his work at YC.