Add files via upload

Author: adoresever

README.md

@@ -0,0 +1,173 @@
+# DateGraphX (Learning Edition)
+
+[English](#english) | [中文](#chinese)
+
+> ⚠️ **Note**: This is a learning edition. For commercial use, please contact us for customized solutions!
+> 
+> ⚠️ **注意**: 这是学习版本。商业用途请联系我们定制解决方案！
+
+<a name="english"></a>
+## 🌟 DateGraphX
+
+An intelligent document analysis system that combines LangChain, Neo4j graph database, and large language models to create a knowledge graph-based RAG (Retrieval-Augmented Generation) application.
+
+### 🖼️ Project Demo
+
+#### Q&A System Interface
+![Q&A System](qa.jpg)
+
+#### Knowledge Graph Visualization
+![Knowledge Graph](kg.jpg)
+
+### 🚀 Features
+
+- 📊 Automatic Knowledge Graph Construction
+  - PDF document processing and analysis
+  - Intelligent text segmentation
+  - Relationship extraction
+  - Interactive graph visualization
+
+- 🤖 Natural Language Q&A
+  - Context-aware responses
+  - Knowledge graph-based retrieval
+  - Multi-LLM support (DeepSeek, OpenAI)
+  - Real-time graph exploration
+
+### 📦 Project Structure
+```
+DateGraphX/
+├── app.py                      # Main application file
+├── api_utils.py                # API utilities
+├── config.py                   # Configuration settings
+├── data_persistence_utils.py   # Data persistence helpers
+├── knowledge_graph_utils.py    # Knowledge graph functions
+├── requirements.txt            # Project dependencies
+├── cache/                      # Cache directory
+├── logo.png                    # Project logo
+├── kg.jpg                      # Knowledge graph demo
+└── qa.jpg                      # Q&A interface demo
+```
+
+### 🔧 Installation
+
+1. Clone repository:
+```bash
+git clone https://github.com/adoresever/DateGraphX.git
+cd DateGraphX
+```
+
+2. Create and activate conda environment:
+```bash
+conda create -n datagraphx python=3.10
+conda activate datagraphx
+```
+
+3. Install dependencies:
+```bash
+pip install -r requirements.txt
+```
+
+4. Start application:
+```bash
+streamlit run app.py
+```
+
+### 🛠️ Requirements
+
+- Python 3.10+
+- Neo4j Database Server
+- DeepSeek/OpenAI API access
+- CUDA-compatible GPU (recommended)
+
+---
+
+<a name="chinese"></a>
+## 🌟 DateGraphX 学习版
+
+一个智能文档分析系统，结合了 LangChain、Neo4j 图数据库和大型语言模型，创建了一个基于知识图谱的 RAG（检索增强生成）应用。
+
+### 🖼️ 项目展示
+
+#### 知识图谱可视化
+![知识图谱](kg.jpg)
+
+#### 问答系统界面
+![问答系统](qa.jpg)
+
+### 🚀 功能特点
+
+- 📊 自动知识图谱构建
+  - PDF文档处理与分析
+  - 智能文本分段
+  - 关系抽取
+  - 交互式图谱可视化
+
+- 🤖 自然语言问答
+  - 上下文感知响应
+  - 基于知识图谱的检索
+  - 多LLM支持（DeepSeek、OpenAI）
+  - 实时图谱探索
+
+### 📦 项目结构
+```
+DateGraphX/
+├── app.py                      # 主应用程序文件
+├── api_utils.py                # API工具
+├── config.py                   # 配置设置
+├── data_persistence_utils.py   # 数据持久化助手
+├── knowledge_graph_utils.py    # 知识图谱功能
+├── requirements.txt            # 项目依赖
+├── cache/                      # 缓存目录
+├── logo.png                    # 项目标志
+├── kg.jpg                      # 知识图谱演示
+└── qa.jpg                      # 问答界面演示
+```
+
+### 🔧 安装步骤
+
+1. 克隆仓库：
+```bash
+git clone https://github.com/adoresever/DateGraphX.git
+cd DateGraphX
+```
+
+2. 创建并激活conda环境：
+```bash
+conda create -n datagraphx python=3.10
+conda activate datagraphx
+```
+
+3. 安装依赖：
+```bash
+pip install -r requirements.txt
+```
+
+4. 启动应用：
+```bash
+streamlit run app.py
+```
+
+### 🛠️ 环境要求
+
+- Python 3.10+
+- Neo4j 数据库服务器
+- DeepSeek/OpenAI API 访问权限
+- CUDA兼容GPU（推荐）
+
+## 👥 作者
+
+**王宇** (Yu Wang) - [[email protected]](mailto:[email protected])
+
+## 📝 致谢
+
+本项目的知识图谱部分参考了 [LightRAG](https://github.com/HKUDS/LightRAG)
+
+## 📄 许可证
+
+CC BY-NC-SA 4.0 - 详见 [LICENSE](LICENSE) 文件
+
+---
+
+> 🔒 **商业定制**
+> 
+> 如需商业版本或定制开发，请联系：[[email protected]](mailto:[email protected])

api_utils.py

@@ -0,0 +1,70 @@
+# api_utils.py
+
+from typing import List, Tuple
+import requests
+from openai import OpenAI
+from langchain_community.embeddings import OpenAIEmbeddings
+from langchain.embeddings.base import Embeddings
+from config import API_CONFIG, EMBEDDING_CONFIG
+
+class LocalEmbeddings(Embeddings):
+    def __init__(self, base_url: str, model: str):
+        self.base_url = base_url
+        self.model = model
+
+    def embed_documents(self, texts: List[str]) -> List[List[float]]:
+        url = f"{self.base_url}/embeddings"
+        embeddings = []
+        for text in texts:
+            response = requests.post(url, json={"input": text, "model": self.model})
+            embeddings.append(response.json()["data"][0]["embedding"])
+        return embeddings
+
+    def embed_query(self, text: str) -> List[float]:
+        url = f"{self.base_url}/embeddings"
+        response = requests.post(url, json={"input": text, "model": self.model})
+        return response.json()["data"][0]["embedding"]
+
+def clean_api_response(response: str, api_type: str) -> str:
+    """清理API响应"""
+    if api_type == "DeepSeek":
+        return response.replace("<｜end▁of▁sentence｜>", "").strip()
+    return response.strip()
+
+def test_api_connection(api_type: str, api_key: str, model_name: str) -> Tuple[bool, str]:
+    """测试API连接"""
+    try:
+        base_url = API_CONFIG[api_type.lower()]['base_url']
+        client = OpenAI(api_key=api_key, base_url=base_url)
+        
+        response = client.chat.completions.create(
+            model=model_name,
+            messages=[{"role": "user", "content": "Hi!"}],
+            max_tokens=10
+        )
+        raw_response = response.choices[0].message.content
+        cleaned_response = clean_api_response(raw_response, api_type)
+        return True, cleaned_response
+    except Exception as e:
+        return False, str(e)
+
+def test_embeddings(embed_type: str, api_key: str = None, base_url: str = None, model: str = None) -> Tuple[bool, str]:
+    """测试嵌入模型"""
+    try:
+        if embed_type == "本地":
+            embeddings = LocalEmbeddings(
+                base_url=base_url or EMBEDDING_CONFIG['local']['base_url'],
+                model=model or EMBEDDING_CONFIG['local']['model']
+            )
+        else:
+            embeddings = OpenAIEmbeddings(api_key=api_key)
+        
+        test_embedding = embeddings.embed_query("test")
+        return True, f"成功生成嵌入向量，维度: {len(test_embedding)}"
+    except Exception as e:
+        return False, str(e)
+
+def get_api_client(api_type: str, api_key: str, model_name: str) -> OpenAI:
+    """获取API客户端"""
+    base_url = API_CONFIG[api_type.lower()]['base_url']
+    return OpenAI(api_key=api_key, base_url=base_url)