External Datasets Service

English | 中文

English

Project Description

Dify External Datasets Service is a service that implements Dify's external knowledge base API, integrating with RagFlow knowledge base.

Features

• Document retrieval based on knowledge base ID
• Configurable retrieval parameters (top_k, score_threshold)
• Bearer token authentication
• Error handling and standardized response format

Prerequisites

1. Knowledge base created in RagFlow
2. API KEY generated in RagFlow
3. RagFlow API server address available

Docker Compose Deployment

1. Create .env file
2. Configure environment variables

   # RagFlow API server address
   API_URL=http://127.0.0.1:12180

3. Start the service using docker-compose

   docker-compose up -d

4. The service will be available at URL_ADDRESS:18121

Source Code Deployment

1. Clone the repository
2. Create .env file
3. Configure environment variables

   # RagFlow API server address
   API_URL=http://127.0.0.1:12180

4. Start the service

   go run main.go

5. The service will be available at URL_ADDRESS:8080

API Usage

Retrieval Endpoint

• URL: /retrieval

• Method: POST

• Headers:

  • Content-Type: application/json
  • Authorization: Bearer

• Request Body:

{
    "knowledge_id": "your-knowledge-base-id",
    "query": "your search query",
    "retrieval_setting": {
        "top_k": 5,
        "score_threshold": 0.7
    }
}

• Response:

{
    "records": [
        {
            "metadata": {
                "path": "document-url",
                "description": "document-name"
            },
            "score": 0.85,
            "title": "document-title",
            "content": "matched-content"
        }
    ]
}

Adding External Knowledge Bases to Dify

1. Connect to External Knowledge Bases
2. How to get knowledge_id
   • Open the developer tools in RagFlow
   • Open the request interface (/v1/kb/list?page=1&page_size=30&keywords=)
   • View knowledge_id

中文

项目描述

Dify External Datasets Service 是一个实现了 Dify 外部知识库 API 的服务，实现了接入 RagFlow 知识库。

功能特性

• 基于知识库 ID 的文档检索
• 可配置的检索参数（top_k, score_threshold）
• Bearer token 认证
• 错误处理和标准化响应格式

使用前准备

1. RagFlow 中已创建知识库
2. RagFlow 中已创建了API KEY
3. 查看 RagFlow Api 服务器地址

docker-compose 部署

1. 创建 .env 文件
2. 配置环境变量

   # RagFlow Api 服务器地址
   API_URL=http://127.0.0.1:12180

3. 使用 docker-compose 启动服务

   docker-compose up -d

4. 服务将在 URL_ADDRESS:18121 上可用

源码启动服务

1. 克隆代码仓库
2. 创建 .env 文件
3. 配置环境变量

   # RagFlow Api 服务器地址
   API_URL=http://127.0.0.1:12180

4. 启动服务

   go run main.go

5. 服务将在 URL_ADDRESS:8080 上可用

API 使用说明

检索接口

• 接口地址：/retrieval

• 请求方法：POST

• 请求头：

  • Content-Type: application/json
  • Authorization: Bearer <your-token>

• 请求体：

{
    "knowledge_id": "知识库ID",
    "query": "检索查询语句",
    "retrieval_setting": {
        "top_k": 5,
        "score_threshold": 0.7
    }
}

• 响应示例：

{
    "records": [
        {
            "metadata": {
                "path": "文档URL",
                "description": "文档名称"
            },
            "score": 0.85,
            "title": "文档标题",
            "content": "匹配内容"
        }
    ]
}

Dify 中添加外部知识库

1. 连接外部知识库 见文档： 连接外部知识库
2. 如何获取 knowledge_id
   • 在 RagFlow 中打开开发人员工具
   • 在请求接口（/v1/kb/list?page=1&page_size=30&keywords=）中查看 knowledge_id。