如何加载网页
本指南涵盖了如何将网页加载到我们在下游使用的LangChain Document 格式中。网页包含文本、图像和其他多媒体元素，通常用HTML表示。它们可能包含指向其他页面或资源的链接。
LangChain 集成了多种适合网页的解析器。选择合适的解析器取决于您的需求。下面我们展示两种可能性：
- Simple and fast 解析，其中我们为每个网页恢复一个
Document，其内容表示为“扁平化”的字符串； - Advanced 解析，其中我们恢复每个页面的多个
Document对象，允许识别和遍历部分、链接、表格和其他结构。
设置
对于“简单快速”的解析，我们将需要langchain-community和beautifulsoup4库：
%pip install -qU langchain-community beautifulsoup4
对于高级解析，我们将使用 langchain-unstructured：
%pip install -qU langchain-unstructured
简单快速的文本提取
如果您正在寻找嵌入网页中的文本的简单字符串表示，以下方法是合适的。它将返回一个Document对象列表——每个页面一个——包含页面文本的单个字符串。在底层，它使用了beautifulsoup4 Python库。
LangChain 文档加载器实现了 lazy_load 及其异步变体 alazy_load，它们返回 Document objects 的迭代器。我们将在下面使用这些功能。
import bs4
from langchain_community.document_loaders import WebBaseLoader
page_url = "https://python.langchain.com/docs/how_to/chatbots_memory/"
loader = WebBaseLoader(web_paths=[page_url])
docs = []
async for doc in loader.alazy_load():
docs.append(doc)
assert len(docs) == 1
doc = docs[0]
API Reference:WebBaseLoader
USER_AGENT environment variable not set, consider setting it to identify your requests.
print(f"{doc.metadata}\n")
print(doc.page_content[:500].strip())
{'source': 'https://python.langchain.com/docs/how_to/chatbots_memory/', 'title': 'How to add memory to chatbots | \uf8ffü¶úÔ∏è\uf8ffüîó LangChain', 'description': 'A key feature of chatbots is their ability to use content of previous conversation turns as context. This state management can take several forms, including:', 'language': 'en'}
How to add memory to chatbots | ü¶úÔ∏èüîó LangChain
Skip to main contentShare your thoughts on AI agents. Take the 3-min survey.IntegrationsAPI ReferenceMoreContributingPeopleLangSmithLangGraphLangChain HubLangChain JS/TSv0.3v0.3v0.2v0.1üí¨SearchIntroductionTutorialsBuild a Question Answering application over a Graph DatabaseTutorialsBuild a Simple LLM Application with LCELBuild a Query Analysis SystemBuild a ChatbotConversational RAGBuild an Extraction ChainBuild an AgentTaggingd
这本质上是页面HTML中文本的转储。它可能包含诸如标题和导航栏之类的无关信息。如果您熟悉预期的HTML，可以通过BeautifulSoup指定所需的类和其他参数。下面我们只解析文章的正文文本：
loader = WebBaseLoader(
web_paths=[page_url],
bs_kwargs={
"parse_only": bs4.SoupStrainer(class_="theme-doc-markdown markdown"),
},
bs_get_text_kwargs={"separator": " | ", "strip": True},
)
docs = []
async for doc in loader.alazy_load():
docs.append(doc)
assert len(docs) == 1
doc = docs[0]
print(f"{doc.metadata}\n")
print(doc.page_content[:500])
{'source': 'https://python.langchain.com/docs/how_to/chatbots_memory/'}
How to add memory to chatbots | A key feature of chatbots is their ability to use content of previous conversation turns as context. This state management can take several forms, including: | Simply stuffing previous messages into a chat model prompt. | The above, but trimming old messages to reduce the amount of distracting information the model has to deal with. | More complex modifications like synthesizing summaries for long running conversations. | We'll go into more detail on a few techniq
print(doc.page_content[-500:])
a greeting. Nemo then asks the AI how it is doing, and the AI responds that it is fine.'), | HumanMessage(content='What did I say my name was?'), | AIMessage(content='You introduced yourself as Nemo. How can I assist you today, Nemo?')] | Note that invoking the chain again will generate another summary generated from the initial summary plus new messages and so on. You could also design a hybrid approach where a certain number of messages are retained in chat history while others are summarized.
请注意，这需要提前了解正文文本在底层HTML中是如何表示的技术知识。
我们可以使用各种设置来参数化WebBaseLoader，允许指定请求头、速率限制、解析器以及BeautifulSoup的其他kwargs。详情请参阅其API参考。
高级解析
如果我们希望对页面内容进行更细粒度的控制或处理，这种方法非常合适。下面，我们不是为每个页面生成一个Document并通过BeautifulSoup控制其内容，而是生成多个Document对象，表示页面上的不同结构。这些结构可以包括章节标题及其对应的正文文本、列表或枚举、表格等。
在底层，它使用了langchain-unstructured库。有关在LangChain中使用Unstructured的更多信息，请参阅集成文档。
from langchain_unstructured import UnstructuredLoader
page_url = "https://python.langchain.com/docs/how_to/chatbots_memory/"
loader = UnstructuredLoader(web_url=page_url)
docs = []
async for doc in loader.alazy_load():
docs.append(doc)
API Reference:UnstructuredLoader
INFO: Note: NumExpr detected 12 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8.
INFO: NumExpr defaulting to 8 threads.
请注意，在没有任何页面HTML结构预先知识的情况下，我们恢复了正文的自然组织：
for doc in docs[:5]:
print(doc.page_content)
How to add memory to chatbots
A key feature of chatbots is their ability to use content of previous conversation turns as context. This state management can take several forms, including:
Simply stuffing previous messages into a chat model prompt.
The above, but trimming old messages to reduce the amount of distracting information the model has to deal with.
More complex modifications like synthesizing summaries for long running conversations.
ERROR! Session/line number was not unique in database. History logging moved to new session 2747
从特定部分提取内容
每个Document对象代表页面的一个元素。其元数据包含有用的信息，例如其类别：
for doc in docs[:5]:
print(f'{doc.metadata["category"]}: {doc.page_content}')
Title: How to add memory to chatbots
NarrativeText: A key feature of chatbots is their ability to use content of previous conversation turns as context. This state management can take several forms, including:
ListItem: Simply stuffing previous messages into a chat model prompt.
ListItem: The above, but trimming old messages to reduce the amount of distracting information the model has to deal with.
ListItem: More complex modifications like synthesizing summaries for long running conversations.
元素之间也可能存在父子关系——例如，一个段落可能属于一个带有标题的部分。如果某个部分特别重要（例如，用于索引），我们可以隔离相应的Document对象。
例如，下面我们加载了两个网页的“设置”部分的内容：
from typing import List
from langchain_core.documents import Document
async def _get_setup_docs_from_url(url: str) -> List[Document]:
loader = UnstructuredLoader(web_url=url)
setup_docs = []
parent_id = -1
async for doc in loader.alazy_load():
if doc.metadata["category"] == "Title" and doc.page_content.startswith("Setup"):
parent_id = doc.metadata["element_id"]
if doc.metadata.get("parent_id") == parent_id:
setup_docs.append(doc)
return setup_docs
page_urls = [
"https://python.langchain.com/docs/how_to/chatbots_memory/",
"https://python.langchain.com/docs/how_to/chatbots_tools/",
]
setup_docs = []
for url in page_urls:
page_setup_docs = await _get_setup_docs_from_url(url)
setup_docs.extend(page_setup_docs)
API Reference:Document
from collections import defaultdict
setup_text = defaultdict(str)
for doc in setup_docs:
url = doc.metadata["url"]
setup_text[url] += f"{doc.page_content}\n"
dict(setup_text)
{'https://python.langchain.com/docs/how_to/chatbots_memory/': "You'll need to install a few packages, and have your OpenAI API key set as an environment variable named OPENAI_API_KEY:\n%pip install --upgrade --quiet langchain langchain-openai\n\n# Set env var OPENAI_API_KEY or load from a .env file:\nimport dotenv\n\ndotenv.load_dotenv()\n[33mWARNING: You are using pip version 22.0.4; however, version 23.3.2 is available.\nYou should consider upgrading via the '/Users/jacoblee/.pyenv/versions/3.10.5/bin/python -m pip install --upgrade pip' command.[0m[33m\n[0mNote: you may need to restart the kernel to use updated packages.\n",
'https://python.langchain.com/docs/how_to/chatbots_tools/': "For this guide, we'll be using a tool calling agent with a single tool for searching the web. The default will be powered by Tavily, but you can switch it out for any similar tool. The rest of this section will assume you're using Tavily.\nYou'll need to sign up for an account on the Tavily website, and install the following packages:\n%pip install --upgrade --quiet langchain-community langchain-openai tavily-python\n\n# Set env var OPENAI_API_KEY or load from a .env file:\nimport dotenv\n\ndotenv.load_dotenv()\nYou will also need your OpenAI key set as OPENAI_API_KEY and your Tavily API key set as TAVILY_API_KEY.\n"}
页面内容上的向量搜索
一旦我们将页面内容加载到LangChain Document对象中，我们就可以以通常的方式对它们进行索引（例如，用于RAG应用程序）。下面我们使用OpenAI embeddings，尽管任何LangChain嵌入模型都可以胜任。
%pip install -qU langchain-openai
import getpass
import os
if "OPENAI_API_KEY" not in os.environ:
os.environ["OPENAI_API_KEY"] = getpass.getpass("OpenAI API Key:")
from langchain_core.vectorstores import InMemoryVectorStore
from langchain_openai import OpenAIEmbeddings
vector_store = InMemoryVectorStore.from_documents(setup_docs, OpenAIEmbeddings())
retrieved_docs = vector_store.similarity_search("Install Tavily", k=2)
for doc in retrieved_docs:
print(f'Page {doc.metadata["url"]}: {doc.page_content[:300]}\n')
API Reference:InMemoryVectorStore | OpenAIEmbeddings
INFO: HTTP Request: POST https://api.openai.com/v1/embeddings "HTTP/1.1 200 OK"
INFO: HTTP Request: POST https://api.openai.com/v1/embeddings "HTTP/1.1 200 OK"
``````output
Page https://python.langchain.com/docs/how_to/chatbots_tools/: You'll need to sign up for an account on the Tavily website, and install the following packages:
Page https://python.langchain.com/docs/how_to/chatbots_tools/: For this guide, we'll be using a tool calling agent with a single tool for searching the web. The default will be powered by Tavily, but you can switch it out for any similar tool. The rest of this section will assume you're using Tavily.
其他网页加载器
有关可用的LangChain网页加载器列表，请参阅此表。