github/DB-GPT
1
0
Fork 0
mirror of https://github.com/csunny/DB-GPT.git synced 2026-07-17 01:58:47 +00:00
Code Issues Packages Projects Releases Wiki Activity

chore:update readme

Browse Source
This commit is contained in:
aries_ckt
2026-03-20 23:56:24 +08:00
parent 07c498d0f7
commit 1c5c07c32c
10 changed files with 11 additions and 222 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
README.hi.md
View File

@@ -1,7 +1,7 @@
<img src="./assets/LOGO_SMALL.png" alt="Logo" style="vertical-align: middle; height: 24px;" /> DB-GPT: AI Native Data App Development framework with AWEL and Agents
<p align="left">
<img src="./assets/Twitter_LOGO.png" width="100%" />
<img src="./assets/dbgpt_vision.png" width="100%" />
</p>
<div align="center">

5
README.ja.md
View File

@@ -2,7 +2,7 @@
<p align="left">
<img src="./assets/Twitter_LOGO.png" width="100%" />
<img src="./assets/dbgpt_vision.png" width="100%" />
</p>
<div align="center">
@@ -143,14 +143,13 @@ SQLとコードを生成してデータをクエリし、データセットを
### 3. マルチソースデータアクセス
構造化データと非構造化データの両方で動作し、データベース、スプレッドシート、ドキュメント、ナレッジベースが含まれます。
![datasource](./assets/datasources.png)
### 4. スキル駆動の拡張性
ドメイン知識、分析方法、実行ワークフローを再利用可能なスキルとしてパッケージ化します。
![import_github_skill](https://github.com/user-attachments/assets/39f39c36-a014-4a2e-8e14-b3af3f1d2f1c)
![agent_browse_use](https://github.com/user-attachments/assets/21864e9f-2179-4f6f-910f-18463ec2b46e)
### 5. サンドボックス実行
分離された環境でコードとツールを実行して、より安全で可靠性の高い分析を実現します。
![sandbox](https://github.com/user-attachments/assets/bfbd78e0-15e2-42ac-876f-5b91847aadc1)

2
README.ma.md
View File

@@ -2,7 +2,7 @@
<p align="left">
<img src="./assets/Twitter_LOGO.png" width="100%" />
<img src="./assets/dbgpt_vision.png" width="100%" />
</p>
<div align="center">

27
README.md
View File

@@ -1,7 +1,7 @@
# <img src="./assets/LOGO_SMALL.png" alt="Logo" style="vertical-align: middle; height: 24px;" /> DB-GPT: Open-Source Agentic AI Data Assistant
<p align="left">
<img src="./assets/Twitter_LOGO.png" width="100%" />
<img src="./assets/dbgpt_vision.png" width="100%" />
</p>
<div align="center">
@@ -82,6 +82,8 @@ Generate SQL and code to query data, clean datasets, compute metrics, and produc
### 3. Multi-source data access
Work across structured and unstructured sources, including databases, spreadsheets, documents, and knowledge bases.
![datasource](./assets/datasources.png)
### 4. Skills-driven extensibility
Package domain knowledge, analysis methods, and execution workflows into reusable skills.
@@ -91,7 +93,6 @@ Package domain knowledge, analysis methods, and execution workflows into reusabl
![import_github_skill](https://github.com/user-attachments/assets/39f39c36-a014-4a2e-8e14-b3af3f1d2f1c)
![agent_browse_use](https://github.com/user-attachments/assets/21864e9f-2179-4f6f-910f-18463ec2b46e)
### 5. Sandboxed execution
Run code and tools in isolated environments for safer, more reliable analysis.
![sandbox](https://github.com/user-attachments/assets/bfbd78e0-15e2-42ac-876f-5b91847aadc1)
@@ -250,24 +251,6 @@ For Docker, local GPU models (vLLM, llama.cpp), or manual source-code setup, see
- controlled tool use
- reproducible outputs and artifacts
## Platform & Ecosystem
DB-GPT is also a platform for building AI-native data systems.
- **AWEL** for agentic workflow orchestration
- **Agents** for autonomous task execution
- **RAG** for knowledge-enhanced reasoning
- **SMMF** for multi-model management
- **DB-GPT-Hub** for Text2SQL and finetuning workflows
- **dbgpts** for apps, workflows, operators, and templates
- [DB-GPT-Plugins](https://github.com/eosphoros-ai/DB-GPT-Plugins) for plugin-based extension
- [GPT-Vis](https://github.com/eosphoros-ai/GPT-Vis) for visualization protocols
#### DeepWiki
- [DB-GPT](https://deepwiki.com/eosphoros-ai/DB-GPT)
- [DB-GPT-HUB](https://deepwiki.com/eosphoros-ai/DB-GPT-Hub)
- [dbgpts](https://deepwiki.com/eosphoros-ai/dbgpts)
#### Text2SQL Finetune
| LLM | Supported |
@@ -418,10 +401,6 @@ The next generation of **AI + Data** products will be:
DB-GPT aims to help developers and enterprises build that future.
## Image
🌐 [AutoDL Image](https://www.codewithgpu.com/i/eosphoros-ai/DB-GPT/dbgpt)
## Contribution

4
README.ta.md
View File

@@ -2,7 +2,7 @@
<p align="left">
<img src="./assets/Twitter_LOGO.png" width="100%" />
<img src="./assets/dbgpt_vision.png" width="100%" />
</p>
<div align="center">
@@ -317,8 +317,6 @@ LLaMA/LLaMA2, Baichuan, ChatGLM, Wenxin, Tongyi, Zhipu மற்றும் ப
- ஆதரவு தரவுமூலங்கள்
- [தரவுமூலங்கள்](http://docs.dbgpt.cn/docs/modules/connections)
## படம்
🌐 [AutoDL படம்](https://www.codewithgpu.com/i/eosphoros-ai/DB-GPT/dbgpt)
## பங்களிப்பு

19
README.zh.md
View File

@@ -1,7 +1,7 @@
# <img src="./assets/LOGO_SMALL.png" alt="Logo" style="vertical-align: middle; height: 24px;" /> DB-GPT：开源 Agentic AI 数据分析智能助手
<p align="left">
<img src="./assets/Twitter_LOGO.png" width="100%" />
<img src="./assets/dbgpt_vision.png" width="100%" />
</p>
@@ -79,14 +79,13 @@ DB-GPT 不只是一个助手界面，它同时也是一个平台，用于构建
### 3. 多数据源分析
同时处理结构化与非结构化数据，包括数据库、表格文件、文档和知识库。
![datasource](./assets/datasources.png)
### 4. Skills 驱动的可扩展能力
将领域知识、分析方法和执行流程沉淀为 skills，实现复用与扩展。
![import_github_skill](https://github.com/user-attachments/assets/39f39c36-a014-4a2e-8e14-b3af3f1d2f1c)
![agent_browse_use](https://github.com/user-attachments/assets/21864e9f-2179-4f6f-910f-18463ec2b46e)
### 5. 沙箱安全执行
在隔离环境中运行代码和工具，让分析过程更安全、更可控。
![sandbox](https://github.com/user-attachments/assets/bfbd78e0-15e2-42ac-876f-5b91847aadc1)
@@ -250,18 +249,6 @@ cd ~/.dbgpt/DB-GPT && uv run dbgpt start webserver --config ~/.dbgpt/configs/<pr
- 可控工具调用
- 可复现的分析产物与 artifacts
## 平台与生态
DB-GPT 同时也是一个构建 AI Native 数据产品的平台，提供：
- **AWEL**：用于 agentic workflow 编排
- **Agents**：用于自主任务执行
- **RAG**：用于知识增强推理
- **SMMF**：用于多模型管理
- **DB-GPT-Hub**：用于 Text2SQL / 微调工作流
- **dbgpts**：用于应用、工作流、算子与模板生态
- [DB-GPT-Plugins](https://github.com/eosphoros-ai/DB-GPT-Plugins)：插件扩展能力
- [GPT-Vis](https://github.com/eosphoros-ai/DB-GPT-Web)：可视化协议
#### DeepWiki
- [DB-GPT](https://deepwiki.com/eosphoros-ai/DB-GPT)
@@ -434,8 +421,6 @@ DB-GPT 希望帮助开发者与企业共同构建这样的未来。
## Image
🌐 [AutoDL镜像](https://www.codewithgpu.com/i/eosphoros-ai/DB-GPT/dbgpt)
🌐 [小程序云部署](https://www.yuque.com/eosphoros/dbgpt-docs/ek12ly8k661tbyn8)
## 使用说明

2
READMR.hi.md
View File

@@ -1,7 +1,7 @@
<img src="./assets/LOGO_SMALL.png" alt="Logo" style="vertical-align: middle; height: 24px;" /> DB-GPT: AI Native Data App Development framework with AWEL and Agents
<p align="left">
<img src="./assets/Twitter_LOGO.png" width="100%" />
<img src="./assets/dbgpt_vision.png" width="100%" />
</p>
<div align="center">

BIN
assets/datasources.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 656 KiB

BIN
assets/dbgpt_vision.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 5.8 MiB

172
docs/docs/getting-started/concepts/architecture.md
View File

@@ -1,172 +0,0 @@
---
sidebar_position: 0
title: Architecture
summary: "High-level map of DB-GPT packages, subsystems, and request flow"
read_when:
- You want to understand how DB-GPT is split across packages and runtime layers
- You need the shortest mental model before diving into AWEL, agents, or RAG
---
# Architecture
An overview of how DB-GPT is structured and how its components fit together.
## High-level view
```mermaid
flowchart TB
subgraph Client["Client Layer"]
WebUI["Web UI"]
CLI["CLI"]
API["REST API"]
end
subgraph App["Application Layer"]
AppMgr["App Manager"]
ChatMgr["Chat Manager"]
FlowMgr["AWEL Flow Manager"]
end
subgraph Core["Core Layer"]
Agent["Agent Framework"]
AWEL["AWEL Engine"]
RAG["RAG Framework"]
SMMF["SMMF (Model Management)"]
end
subgraph Data["Data Layer"]
DS["Data Sources"]
KB["Knowledge Base"]
VectorDB["Vector Store"]
MetaDB["Metadata Store"]
end
Client --> App
App --> Core
Core --> Data
SMMF --> LLM["LLM Providers"]
```
## Package structure
DB-GPT is a Python monorepo organized into multiple packages under `packages/`:
| Package | Purpose |
|---|---|
| **dbgpt-core** | Core abstractions: agent, AWEL, RAG, model interfaces, storage |
| **dbgpt-app** | Application server, chat logic, Web API endpoints |
| **dbgpt-serve** | Service modules (knowledge, flow, app, datasource management) |
| **dbgpt-ext** | Extensions: datasource connectors, storage backends, model providers |
| **dbgpt-client** | Python client SDK for the DB-GPT REST API |
| **dbgpt-accelerator** | GPU acceleration utilities (quantization, inference optimization) |
```
DB-GPT/
├── packages/
│ ├── dbgpt-core/ # Core abstractions
│ ├── dbgpt-app/ # Application server
│ ├── dbgpt-serve/ # Service modules
│ ├── dbgpt-ext/ # Extensions
│ ├── dbgpt-client/ # Python client SDK
│ └── dbgpt-accelerator/ # GPU acceleration
├── web/ # Next.js Web UI
├── configs/ # TOML configuration files
└── docs/ # Documentation (Docusaurus)
```
## Core subsystems
### SMMF (Service-oriented Multi-Model Management Framework)
Manages multiple LLM and embedding model instances. Supports:
- API proxy models (OpenAI, DeepSeek, Qwen, etc.)
- Local models via HuggingFace Transformers, vLLM, llama.cpp
- Model switching and failover
- Standalone and cluster deployment modes
Learn more: [SMMF Concept](/docs/getting-started/concepts/smmf) | [SMMF Module](/docs/modules/smmf)
### AWEL (Agentic Workflow Expression Language)
A domain-specific language for building AI application workflows as directed acyclic graphs (DAGs). AWEL provides:
- Operators: Map, Reduce, Join, Branch, Stream transformers
- Triggers: HTTP, scheduler-based
- Visual editor: AWEL Flow in the Web UI
Learn more: [AWEL Concept](/docs/getting-started/concepts/awel) | [AWEL Tutorial](/docs/awel/tutorial)
### Agent Framework
Data-driven multi-agent system with:
- **Profile**: Agent identity and role definition
- **Memory**: Sensory, short-term, long-term, and hybrid memory
- **Planning**: Task decomposition and execution strategies
- **Action**: Tool invocation and result processing
- **Resource**: Tools, databases, knowledge bases, and resource packs
Learn more: [Agents Concept](/docs/getting-started/concepts/agents) | [Agent Guide](/docs/agents/introduction/)
### RAG Framework
Retrieval-Augmented Generation with multiple retrieval strategies:
- Vector similarity search (ChromaDB, Milvus, OceanBase)
- Knowledge graph retrieval (Graph RAG)
- Keyword-based retrieval (BM25)
- Hybrid retrieval combining multiple strategies
Learn more: [RAG Concept](/docs/getting-started/concepts/rag) | [RAG Module](/docs/modules/rag)
## Configuration
DB-GPT uses TOML configuration files in the `configs/` directory:
```toml
# configs/dbgpt-proxy-openai.toml
[models]
[[models.llms]]
name = "chatgpt_proxyllm"
provider = "proxy/openai"
api_key = "your-api-key"
[[models.embeddings]]
name = "text-embedding-3-small"
provider = "proxy/openai"
api_key = "your-api-key"
```
Full reference: [Config Reference](/docs/config/config-reference)
## Data flow
A typical chat request flows through DB-GPT like this:
```mermaid
sequenceDiagram
participant User
participant WebUI
participant Server
participant Agent
participant LLM
participant DataSource
User->>WebUI: Send message
WebUI->>Server: POST /api/v2/chat/completions
Server->>Agent: Route to agent
Agent->>DataSource: Query data (if needed)
DataSource-->>Agent: Data results
Agent->>LLM: Generate response
LLM-->>Agent: Model output
Agent-->>Server: Formatted response
Server-->>WebUI: Stream response
WebUI-->>User: Display answer
```
## What's next
- [AWEL](/docs/getting-started/concepts/awel) — Understand workflow orchestration
- [Agents](/docs/getting-started/concepts/agents) — Learn about the agent framework
- [Model Providers](/docs/getting-started/providers/) — Configure your preferred LLM