Apache Doris

Apache Doris is a modern data warehouse for real-time analytics. It delivers lightning-fast analytics on real-time data at scale.

Usually Apache Doris is categorized into OLAP, and it has showed excellent performance in ClickBench — a Benchmark For Analytical DBMS. Since it has a super-fast vectorized execution engine, it could also be used as a fast vectordb.

您需要使用 pip install -qU langchain-community 安装 langchain-community 才能使用此集成

本文将展示如何使用 Apache Doris 向量存储。

设置

%pip install --upgrade --quiet  pymysql

在开头设置 update_vectordb = False。如果没有文档更新，则无需重建文档的嵌入向量

!pip install  sqlalchemy
!pip install langchain

from langchain.chains import RetrievalQA
from langchain_community.document_loaders import (
    DirectoryLoader,
    UnstructuredMarkdownLoader,
)
from langchain_community.vectorstores.apache_doris import (
    ApacheDoris,
    ApacheDorisSettings,
)
from langchain_openai import OpenAI, OpenAIEmbeddings
from langchain_text_splitters import TokenTextSplitter

update_vectordb = False

API 参考：RetrievalQA | DirectoryLoader | UnstructuredMarkdownLoader | ApacheDoris | ApacheDorisSettings | OpenAI | OpenAIEmbeddings | TokenTextSplitter

加载文档并将其分割为令牌

加载 docs 目录下的所有 Markdown 文件

对于 Apache Doris 文档，您可以从 https://github.com/apache/doris 克隆仓库，其中包含一个 docs 目录。

loader = DirectoryLoader(
    "./docs", glob="**/*.md", loader_cls=UnstructuredMarkdownLoader
)
documents = loader.load()

将文档拆分为令牌，并设置 update_vectordb = True，因为存在新的文档/令牌。

# load text splitter and split docs into snippets of text
text_splitter = TokenTextSplitter(chunk_size=400, chunk_overlap=50)
split_docs = text_splitter.split_documents(documents)

# tell vectordb to update text embeddings
update_vectordb = True

split_docs[-20]

print("# docs = %d, # splits = %d" % (len(documents), len(split_docs)))

创建向量数据库实例

使用 Apache Doris 作为向量数据库

def gen_apache_doris(update_vectordb, embeddings, settings):
    if update_vectordb:
        docsearch = ApacheDoris.from_documents(split_docs, embeddings, config=settings)
    else:
        docsearch = ApacheDoris(embeddings, settings)
    return docsearch

将令牌转换为嵌入并放入向量数据库

此处我们使用 Apache Doris 作为向量数据库，您可以通过 ApacheDorisSettings 配置 Apache Doris 实例。

配置 Apache Doris 实例与配置 MySQL 实例非常相似。您需要指定：

1. host/port
2. 用户名（默认：'root'）
3. 密码（默认：''）
4. 数据库（默认值：'default'）
5. 表（默认值：'langchain'）

import os
from getpass import getpass

if "OPENAI_API_KEY" not in os.environ:
    os.environ["OPENAI_API_KEY"] = getpass()

update_vectordb = True

embeddings = OpenAIEmbeddings()

# configure Apache Doris settings(host/port/user/pw/db)
settings = ApacheDorisSettings()
settings.port = 9030
settings.host = "172.30.34.130"
settings.username = "root"
settings.password = ""
settings.database = "langchain"
docsearch = gen_apache_doris(update_vectordb, embeddings, settings)

print(docsearch)

update_vectordb = False

构建问答系统并向其提问

llm = OpenAI()
qa = RetrievalQA.from_chain_type(
    llm=llm, chain_type="stuff", retriever=docsearch.as_retriever()
)
query = "what is apache doris"
resp = qa.run(query)
print(resp)

相关

• 向量存储 概念指南
• 向量存储 操操作指南