探索SAP HANA Cloud Vector Engine：构建高效的自查询系统

最新推荐文章于 2024-10-20 02:37:56 发布

tt_jishu

最新推荐文章于 2024-10-20 02:37:56 发布

阅读量289

点赞数 3

文章标签：数据库服务器前端 python

本文链接：https://blog.csdn.net/tt_jishu/article/details/142291036

版权

# 探索SAP HANA Cloud Vector Engine：构建高效的自查询系统

## 引言

在现代数据驱动的世界中，能够有效地存储和查询向量化数据是非常重要的。SAP HANA Cloud Vector Engine 提供了一种高效的存储和检索解决方案。本文将深入探讨如何使用SAP HANA Cloud Vector Engine 构建一个支持自查询的系统，并展示相关代码示例。

## 主要内容

### 环境设置

在开始之前，请确保你已在SAP HANA上设置好向量存储，并具有必要的连接信息。以下是环境变量设置的示例：

```python
import os
from hdbcli import dbapi

# 使用API代理服务提高访问稳定性
connection = dbapi.connect(
    address=os.environ.get("HANA_DB_ADDRESS"),
    port=os.environ.get("HANA_DB_PORT"),
    user=os.environ.get("HANA_DB_USER"),
    password=os.environ.get("HANA_DB_PASSWORD"),
    autocommit=True,
    sslValidateCertificate=False,
)

创建表并添加数据

首先，我们创建一个自定义表，用于存储文档及其元数据。

# 创建自定义表
cur = connection.cursor()
cur.execute("DROP TABLE LANGCHAIN_DEMO_SELF_QUERY", ignoreErrors=True)
cur.execute(
    """CREATE TABLE "LANGCHAIN_DEMO_SELF_QUERY"  (
    "name" NVARCHAR(100), "is_active" BOOLEAN, "id" INTEGER, "height" DOUBLE,
    "VEC_TEXT" NCLOB, 
    "VEC_META" NCLOB, 
    "VEC_VECTOR" REAL_VECTOR
    )"""
)

接下来，添加一些示例文档。

from langchain_community.vectorstores.hanavector import HanaDB
from langchain_core.documents import Document
from langchain_openai import OpenAIEmbeddings

embeddings = OpenAIEmbeddings()

# 准备测试文档
docs = [
    Document(page_content="First", metadata={"name": "adam", "is_active": True, "id": 1, "height": 10.0}),
    Document(page_content="Second", metadata={"name": "bob", "is_active": False, "id": 2, "height": 5.7}),
    Document(page_content="Third", metadata={"name": "jane", "is_active": True, "id": 3, "height": 2.4}),
]

db = HanaDB(
    connection=connection,
    embedding=embeddings,
    table_name="LANGCHAIN_DEMO_SELF_QUERY",
    specific_metadata_columns=["name", "is_active", "id", "height"],
)

# 删除旧文档并添加新文档
db.delete(filter={})
db.add_documents(docs)

构建自查询系统

使用 SelfQueryRetriever 实现自查询功能。

from langchain.chains.query_constructor.base import AttributeInfo
from langchain.retrievers.self_query.base import SelfQueryRetriever
from langchain_community.query_constructors.hanavector import HanaTranslator
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(model="gpt-3.5-turbo")

metadata_field_info = [
    AttributeInfo(name="name", description="The name of the person", type="string"),
    AttributeInfo(name="is_active", description="Whether the person is active", type="boolean"),
    AttributeInfo(name="id", description="The ID of the person", type="integer"),
    AttributeInfo(name="height", description="The height of the person", type="float"),
]

document_content_description = "A collection of persons"

hana_translator = HanaTranslator()

retriever = SelfQueryRetriever.from_llm(
    llm,
    db,
    document_content_description,
    metadata_field_info,
    structured_query_translator=hana_translator,
)

# 使用检索器进行查询
query_prompt = "Which person is not active?"
docs = retriever.invoke(input=query_prompt)
for doc in docs:
    print("-" * 80)
    print(doc.page_content, " ", doc.metadata)

查看查询构建过程

分析查询的结构化输出。

from langchain.chains.query_constructor.base import (
    StructuredQueryOutputParser,
    get_query_constructor_prompt,
)

prompt = get_query_constructor_prompt(document_content_description, metadata_field_info)
output_parser = StructuredQueryOutputParser.from_components()
query_constructor = prompt | llm | output_parser

sq = query_constructor.invoke(input=query_prompt)

print("Structured query: ", sq)
print("Translated for hana vector store: ", hana_translator.visit_structured_query(sq))