Qdrant - Installation

Ragfish provides Qdrant integration through the @ragfish/qdrant package.

Qdrant acts as the vector store for storing and searching the embeddings generated from your knowledge. Ragfish also provides QdrantRetriever to connect Qdrant with the retrieval layer. The current Ragfish architecture uses both QdrantVectorStore and QdrantRetriever.

Prerequisites

Before using Qdrant with Ragfish, you need:

1. Node.js

2. A Ragfish project

3. A running Qdrant instance

4. The required Qdrant connection details

Qdrant can be deployed according to your application's infrastructure requirements.

Install the Ragfish Qdrant Package

Install the Qdrant integration:

npm install @ragfish/qdrant

Your application should also have the Ragfish Core package installed:

npm install @ragfish/core

If you are using OpenAI for embeddings and response generation:

npm install @ragfish/openai

A typical project therefore uses:

@ragfish/core
@ragfish/openai
@ragfish/qdrant

Qdrant Architecture

The Qdrant integration provides two important components:

@ragfish/qdrant
│
├── QdrantVectorStore
└── QdrantRetriever

Basic Setup

Import the Qdrant components:

import {
 QdrantVectorStore,
 QdrantRetriever
} from "@ragfish/qdrant";
Create a vector store:
const store = new QdrantVectorStore({
 // Qdrant configuration
});
Then create a retriever:
const retriever = new QdrantRetriever({
 vectorStore: store,
 collectionName: "knowledge"
});

The Ragfish master example follows this same pattern, creating a QdrantVectorStore and passing it to QdrantRetriever with a collection name.

Environment Variables

For production applications, keep Qdrant connection information outside your source code.

For example:

QDRANT_URL=your-qdrant-url
QDRANT_API_KEY=your-qdrant-api-key

The exact configuration properties supported by QdrantVectorStore should be taken from the version of @ragfish/qdrant installed in your project.

Qdrant with OpenAI

A typical Ragfish RAG application can combine Qdrant with OpenAI:

OpenAIEmbedding
      │
      ▼
QdrantVectorStore
      │
      ▼
QdrantRetriever
      │
      ▼
     Chat
      │
      ▼
 OpenAILLM

The embedding model generates vectors, Qdrant stores and searches those vectors, the retriever returns relevant content, and the LLM generates the final response.

Complete Setup

A basic application structure looks like:

import { Settings, Chat } from "@ragfish/core";

import {
 OpenAILLM,
 OpenAIEmbedding
} from "@ragfish/openai";

import {
 QdrantVectorStore,
 QdrantRetriever
} from "@ragfish/qdrant";

Settings.llm = new OpenAILLM({
 apiKey: process.env.OPENAI_API_KEY
});

Settings.embedModel = new OpenAIEmbedding({
 apiKey: process.env.OPENAI_API_KEY
});

const store = new QdrantVectorStore({
 // Qdrant configuration
});

const retriever = new QdrantRetriever({
 vectorStore: store,
 collectionName: "knowledge"
});

const chat = new Chat({
 retriever
});

Verify the Installation

After installation, verify that the packages can be imported successfully:

import { Chat } from "@ragfish/core";

import {
 OpenAILLM,
 OpenAIEmbedding
} from "@ragfish/openai";

import {
 QdrantVectorStore,
 QdrantRetriever
} from "@ragfish/qdrant";

If these imports resolve successfully, the required Ragfish packages are available to your project.

Troubleshooting

Package Not Found

If the package cannot be resolved, verify that it is installed:

npm install @ragfish/qdrant

Qdrant Connection Failure

Check:

1. Qdrant URL

2. Authentication credentials

3. Network connectivity

4. Qdrant service availability

Collection Not Found

Make sure the collection configured in QdrantRetriever exists and contains the expected indexed knowledge.

For Example:

const retriever = new QdrantRetriever({
 vectorStore: store,
 collectionName: "knowledge"
});

The collection name must correspond to the knowledge collection you intend to search.