GaussDB for LangChain

An integration package connecting GaussDB and LangChain, supports quick connection to GaussDB and integrates LangChain workflows.

Installation

• the first method: git clone this repo, then install with:

poetry install

• the second method: install with pip (already been released):

pip install langchain_gaussdb-0.1.0-py3-none-any.whl

Quick Start

Initialization

from langchain_openai import OpenAIEmbeddings

embeddings = OpenAIEmbeddings(
    model="Qwen/Qwen3-Embedding-8B",
    api_key="xxx",
    base_url="https://api.siliconflow.cn/v1",
    dimensions=1024
)

from langchain_gaussdb import GaussVectorSettings, GaussVectorStore

config = GaussVectorSettings(
    host="10.25.106.120",
    port=9800,
    user="langchain_gv",
    password="Gauss_234",
    database="postgres",
    table_name="my_docs"
)

vector_store = GaussVectorStore(
    embedding=embeddings,
    config=config
)

Insert documents

We can use the add_documents function to add items to the vector store. Note that adding a document by ID will overwrite any existing document that matches that ID.

from langchain_core.documents import Document

docs = [
    Document(
        page_content="there are fishes in the pond",
        metadata={"id": 1, "location": "pond", "topic": "animals"},
    ),
    Document(
        page_content="ducks are also found in the pond",
        metadata={"id": 2, "location": "pond", "topic": "animals"},
    ),
    Document(
        page_content="tigers are big and strong",
        metadata={"id": 3, "location": "zoo", "topic": "animals"},
    ),
    Document(
        page_content="fresh apples are available at the market",
        metadata={"id": 4, "location": "market", "topic": "food"},
    ),
    Document(
        page_content="the market also sells fresh oranges",
        metadata={"id": 5, "location": "market", "topic": "food"},
    ),
    Document(
        page_content="the new art exhibit is fascinating",
        metadata={"id": 6, "location": "museum", "topic": "art"},
    ),
    Document(
        page_content="a sculpture exhibit is also at the museum",
        metadata={"id": 7, "location": "museum", "topic": "art"},
    ),
    Document(
        page_content="a new coffee shop opened on Main Street",
        metadata={"id": 8, "location": "Main Street", "topic": "food"},
    ),
    Document(
        page_content="the book club meets at the library",
        metadata={"id": 9, "location": "library", "topic": "reading"},
    ),
    Document(
        page_content="the library hosts a weekly story time for kids",
        metadata={"id": 10, "location": "library", "topic": "reading"},
    ),
]

vector_store.add_documents(docs, ids=[doc.metadata["id"] for doc in docs])

Delete documents

vector_store.delete(ids=["9"])

Similarity search

A simple similarity search with metadata filtering can be performed as follows.

results = vector_store.similarity_search(
    "ducks", k=10, filter={"location": "pond", "topic": "animals"}
)
for doc in results:
    print(f"* {doc.page_content} [{doc.metadata}]")

If you want to perform a similarity search and receive the corresponding score, you can run the following command.

results = vector_store.similarity_search_with_score(query="tigers", k=1)
for doc, score in results:
    print(f"* [SIM={score:3f}] {doc.page_content} [{doc.metadata}]")

Retriever

You can also convert vector stores into retrievers for more convenient use within your pipeline.

retriever = vector_store.as_retriever(search_type="similarity", search_kwargs={"k": 1})
results = retriever.invoke("kitty")
for doc in results:
    print(f"* {doc.page_content} [{doc.metadata}]")

Configuration Guide

Connection Settings

Parameter	Default	Description
host	localhost	Database server address
port	5432	Database connection port
user	gaussdb	Database username
password	-	Complex password string
database	postgres	Default database name
min_connections	1	Connection pool minimum size
max_connections	5	Connection pool maximum size
table_name	langchain_docs	Name of the table for storing vector data and metadata
index_type	IndexType.GSDISKANN	Vector index algorithm type. Options: GsIVFFLAT or GsDiskANN. Default is GsDiskANN.
index_params	DiskAnnParams()	Vector index params, different index_type require different index_params.
vector_type	VectorType.float_vector	Type of vector representation to use. Options: floatvector or boolvector. Default is floatvector.
distance_strategy	DistanceStrategy.COSINE	Vector similarity metric to use for retrieval. Options: l2, cosine, hamming. Default is cosine.
embedding_dimension	1024	Dimensionality of the vector embeddings.

Supported Combinations

Index Types	Dimensions	Vector Type	Supported Distance Strategies
GsIVFFLAT	<=1024	floatvector/boolvector	l2/cosine/hamming
GsDiskANN	<=4096	floatvector	l2/cosine

API Reference

Method	Description
add_documents	Insert documents with automatic embedding
similarity_search	Basic vector similarity search
similarity_search_with_score	Return (document, similarity_score) tuples
delete	Remove documents by ID list
get_by_ids	Get documents by ID list
drop_table	Delete entire table
from_documents	Initialize GaussVectorStore from documents