AnalyticDB

AnalyticDB for PostgreSQL is a massively parallel processing (MPP) data warehousing service that is designed to analyze large volumes of data online.

AnalyticDB for PostgreSQL is developed based on the open source Greenplum Database project and is enhanced with in-depth extensions by Alibaba Cloud. AnalyticDB for PostgreSQL is compatible with the ANSI SQL 2003 syntax and the PostgreSQL and Oracle database ecosystems. AnalyticDB for PostgreSQL also supports row store and column store. AnalyticDB for PostgreSQL processes petabytes of data offline at a high performance level and supports highly concurrent online queries.

This notebook shows how to use functionality related to the AnalyticDB vector database.

To run, you should have an AnalyticDB instance up and running:

• Using AnalyticDB Cloud Vector Database.

Compatibility

Only available on Node.js.

Setup

LangChain.js accepts node-postgres as the connections pool for AnalyticDB vectorstore.

npm

npm install -S pg

Yarn

yarn add pg

pnpm

pnpm add pg

And we need pg-copy-streams to add batch vectors quickly.

npm

npm install -S pg-copy-streams

Yarn

yarn add pg-copy-streams

pnpm

pnpm add pg-copy-streams

tip

See this section for general instructions on installing integration packages.

npm

npm install @langchain/openai @langchain/community

Yarn

yarn add @langchain/openai @langchain/community

pnpm

pnpm add @langchain/openai @langchain/community

Usage

Security

User-generated data such as usernames should not be used as input for the collection name.
This may lead to SQL Injection!

import { AnalyticDBVectorStore } from "@langchain/community/vectorstores/analyticdb";
import { OpenAIEmbeddings } from "@langchain/openai";

const connectionOptions = {
  host: process.env.ANALYTICDB_HOST || "localhost",
  port: Number(process.env.ANALYTICDB_PORT) || 5432,
  database: process.env.ANALYTICDB_DATABASE || "your_database",
  user: process.env.ANALYTICDB_USERNAME || "username",
  password: process.env.ANALYTICDB_PASSWORD || "password",
};

const vectorStore = await AnalyticDBVectorStore.fromTexts(
  ["foo", "bar", "baz"],
  [{ page: 1 }, { page: 2 }, { page: 3 }],
  new OpenAIEmbeddings(),
  { connectionOptions }
);
const result = await vectorStore.similaritySearch("foo", 1);
console.log(JSON.stringify(result));
// [{"pageContent":"foo","metadata":{"page":1}}]

await vectorStore.addDocuments([{ pageContent: "foo", metadata: { page: 4 } }]);

const filterResult = await vectorStore.similaritySearch("foo", 1, {
  page: 4,
});
console.log(JSON.stringify(filterResult));
// [{"pageContent":"foo","metadata":{"page":4}}]

const filterWithScoreResult = await vectorStore.similaritySearchWithScore(
  "foo",
  1,
  { page: 3 }
);
console.log(JSON.stringify(filterWithScoreResult));
// [[{"pageContent":"baz","metadata":{"page":3}},0.26075905561447144]]

const filterNoMatchResult = await vectorStore.similaritySearchWithScore(
  "foo",
  1,
  { page: 5 }
);
console.log(JSON.stringify(filterNoMatchResult));
// []

// need to manually close the Connection pool
await vectorStore.end();

API Reference:

• AnalyticDBVectorStore from @langchain/community/vectorstores/analyticdb
• OpenAIEmbeddings from @langchain/openai

Related

• Vector store conceptual guide
• Vector store how-to guides