RAG Chatbot - AI-Powered Document Assistant

A full-stack RAG (Retrieval-Augmented Generation) chatbot that allows users to upload PDF and DOCX documents and chat with an AI about their content. Built with React, Express.js, LangChain, FAISS, and Google Gemini.

🚀 Features

📄 Document Upload: Upload one or multiple PDF and DOCX files (max 5 files, 10MB each)
🤖 AI-Powered Chat: Ask questions about uploaded documents using Google Gemini 2.0 Flash
🔒 Session-Based: Each user session maintains its own vector store and chat history
🧹 Auto-Cleanup: Documents and embeddings automatically cleared on session end
💅 Responsive UI: Modern, clean interface built with React and Tailwind CSS
🐳 Dockerized: Fully containerized with Docker Compose for easy deployment
⚡ Fast Vector Search: FAISS-powered semantic search with HuggingFace embeddings
📝 Markdown Support: Rich markdown rendering in chat responses

📋 Prerequisites

Node.js 20+ (for local development)
Docker and Docker Compose (for containerized deployment)
Google Gemini API Key - Get it from Google AI Studio

🏗️ Architecture

System Architecture Diagram

┌─────────────────────────────────────────────────────────────────┐
│                          CLIENT BROWSER                         │
│                     (React + Tailwind CSS)                      │
└────────────────────────┬────────────────────────────────────────┘
                         │ HTTP/HTTPS
                         │ (Port 80/443)
                         ▼
┌─────────────────────────────────────────────────────────────────┐
│                      NGINX (Frontend)                           │
│              Serves React SPA + Static Assets                   │
└────────────────────────┬────────────────────────────────────────┘
                         │ API Calls
                         │ (Port 5000)
                         ▼
┌─────────────────────────────────────────────────────────────────┐
│                   EXPRESS.JS BACKEND                            │
│  ┌──────────────────────────────────────────────────────────┐   │
│  │  Middleware Layer                                        │   │
│  │  • CORS • Helmet • Sessions • Error Handler • Multer     │   │
│  └──────────────────────────────────────────────────────────┘   │
│  ┌──────────────────────────────────────────────────────────┐   │
│  │  Routes                                                  │   │
│  │  • /api/upload  • /api/chat  • /api/session              │   │
│  └──────────────────────────────────────────────────────────┘   │
│  ┌──────────────────────────────────────────────────────────┐   │
│  │  Controllers                                             │   │
│  │  • uploadController  • chatController  • sessionCtrl     │   │
│  └──────────────────────────────────────────────────────────┘   │
│  ┌──────────────────────────────────────────────────────────┐   │
│  │  Services (Business Logic)                               │   │
│  │  • ragService        • vectorStoreService                │   │
│  │  • sessionService    • documentLoader                    │   │
│  └──────────────────────────────────────────────────────────┘   │
└────────────────┬──────────────────────┬────────────────────────-┘
                 │                      │
                 ▼                      ▼
     ┌────────────────────┐  ┌──────────────────────┐
     │     File System    │  │     Memory Store     │
     │   (uploads/)       │  │  • Vector Stores     │
     │   • PDF Files      │  │  • Sessions          │
     │   • DOCX Files     │  │  • Chat History      │
     └────────────────────┘  └──────────────────────┘
                 │
                 ▼
     ┌────────────────────────────────────────┐
     │           External Services            │
     │  ┌──────────────────┐ ┌──────────────┐ │
     │  │ Google Gemini    │ │ HuggingFace  │ │
     │  │ (LLM - 2.0 Flash)│ │ (Embeddings) │ │
     │  └──────────────────┘ └──────────────┘ │
     └────────────────────────────────────────┘

Data Flow Diagram

1. Document Upload Flow

┌──────────┐
│  User    │
└────┬─────┘
     │ 1. Select/Drop Files (PDF/DOCX)
     ▼
┌─────────────────┐
│  FileUpload     │
│  Component      │
└────┬────────────┘
     │ 2. POST /api/upload (multipart/form-data)
     ▼
┌─────────────────────────┐
│  uploadMiddleware       │
│  (Multer)               │
│  • Validate file type   │
│  • Check file size      │
│  • Save to uploads/     │
└────┬────────────────────┘
     │ 3. Files saved
     ▼
┌─────────────────────────┐
│  uploadController       │
│  • Get file paths       │
└────┬────────────────────┘
     │ 4. Process documents
     ▼
┌─────────────────────────┐
│  ragService             │
│  processDocuments()     │
└────┬────────────────────┘
     │ 5. Load documents
     ▼
┌─────────────────────────┐
│  documentLoader         │
│  • Parse PDF/DOCX       │
│  • Extract text         │
└────┬────────────────────┘
     │ 6. Document chunks
     ▼
┌─────────────────────────┐
│  vectorStoreService     │
│  createVectorStore()    │
└────┬────────────────────┘
     │ 7a. Split into chunks (1000 chars, 200 overlap)
     ▼
┌─────────────────────────┐
│  HuggingFace            │
│  all-MiniLM-L6-v2       │
│  • Generate embeddings  │
└────┬────────────────────┘
     │ 7b. Embeddings (384-dim vectors)
     ▼
┌─────────────────────────┐
│  FAISS                  │
│  • Create index         │
│  • Store vectors        │
└────┬────────────────────┘
     │ 8. Vector store created
     ▼
┌─────────────────────────┐
│  sessionService         │
│  • Store vectorStore    │
│  • Link to session      │
└────┬────────────────────┘
     │ 9. Success response
     ▼
┌─────────────────┐
│  Frontend       │
│  • Show success │
│  • Display files│
└─────────────────┘

2. Chat/Query Flow

┌──────────┐
│  User    │
└────┬─────┘
     │ 1. Type question
     ▼
┌─────────────────┐
│  MessageInput   │
│  Component      │
└────┬────────────┘
     │ 2. POST /api/chat { question }
     ▼
┌─────────────────────────┐
│  chatController         │
│  • Get session ID       │
│  • Extract question     │
└────┬────────────────────┘
     │ 3. Process query
     ▼
┌─────────────────────────┐
│  ragService             │
│  chat()                 │
└────┬────────────────────┘
     │ 4. Get session data
     ▼
┌─────────────────────────┐
│  sessionService         │
│  • Retrieve vectorStore │
│  • Get chat history     │
└────┬────────────────────┘
     │ 5a. Embed question
     ▼
┌─────────────────────────┐
│  HuggingFace            │
│  all-MiniLM-L6-v2       │
│  • Convert to vector    │
└────┬────────────────────┘
     │ 5b. Question vector
     ▼
┌─────────────────────────┐
│  FAISS VectorStore      │
│  similaritySearch()     │
│  • Find top 3 matches   │
└────┬────────────────────┘
     │ 6. Retrieved chunks
     ▼
┌─────────────────────────┐
│  ragService             │
│  • Format context       │
│  • Build prompt         │
└────┬────────────────────┘
     │ 7. Prompt with context
     ▼
┌─────────────────────────┐
│  Google Gemini          │
│  gemini-2.0-flash       │
│  • Generate answer      │
└────┬────────────────────┘
     │ 8. AI response
     ▼
┌─────────────────────────┐
│  ragService             │
│  • Save to history      │
└────┬────────────────────┘
     │ 9. Return answer
     ▼
┌─────────────────┐
│  ChatInterface  │
│  • Display msg  │
│  • Render MD    │
└─────────────────┘

Component Interaction Diagram

Frontend (React)                Backend (Express.js)
═══════════════════            ═══════════════════════

┌─────────────┐                ┌──────────────────┐
│   App.jsx   │────────────────│  Session Mgmt    │
│             │  Session Cookie│  Middleware      │
└──────┬──────┘                └──────────────────┘
       │
       ├─────┬─────────┬────────────┐
       │     │         │            │
       ▼     ▼         ▼            ▼
┌──────────┐ ┌─────┐ ┌──────┐ ┌──────────┐
│FileUpload│ │Chat │ │Msg   │ │Session   │
│Component │ │Intf │ │Input │ │Controls  │
└────┬─────┘ └──┬──┘ └───┬──┘ └────┬─────┘
     │          │        │         │
     │          └────────┴─────────┘
     │                   │
     ▼                   ▼
┌─────────────────────────────────┐
│     api.js (Axios)              │
│  • uploadFiles()                │
│  • sendChatMessage()            │
│  • getChatHistory()             │
│  • clearSession()               │
└────────────┬────────────────────┘
             │ HTTP Requests
             ▼
     ┌───────────────────┐
     │   Routes          │
     │  /api/upload      │◄──── uploadMiddleware (Multer)
     │  /api/chat        │
     │  /api/session     │
     └────────┬──────────┘
              │
     ┌────────┴──────────┐
     │   Controllers     │
     │  uploadController │
     │  chatController   │
     │  sessionCtrl      │
     └────────┬──────────┘
              │
     ┌────────┴──────────┐
     │    Services       │
     │  ragService       │◄─── Google Gemini API
     │  vectorStoreService│◄─── HuggingFace
     │  sessionService   │
     │  documentLoader   │
     └───────────────────┘

RAG Pipeline Diagram

┌────────────────────────────────────────────────────────────────┐
│                    RAG PROCESSING PIPELINE                      │
└────────────────────────────────────────────────────────────────┘

Step 1: DOCUMENT INGESTION
────────────────────────────
┌──────────────┐
│ User Upload  │
│ PDF/DOCX     │
└──────┬───────┘
       │
       ▼
┌─────────────────────┐
│ Document Loaders    │
│ • PDFLoader         │
│ • DocxLoader        │
└──────┬──────────────┘
       │ Raw Text
       ▼

Step 2: TEXT CHUNKING
────────────────────────────
┌─────────────────────────────┐
│ RecursiveCharacterSplitter  │
│ • Chunk Size: 1000 chars    │
│ • Overlap: 200 chars        │
│ • Preserves context         │
└──────┬──────────────────────┘
       │ Text Chunks
       ▼

Step 3: EMBEDDING GENERATION
────────────────────────────
┌─────────────────────────────┐
│ HuggingFace Embeddings      │
│ Model: all-MiniLM-L6-v2     │
│ • Input: Text chunks        │
│ • Output: 384-dim vectors   │
└──────┬──────────────────────┘
       │ Vector Embeddings
       ▼

Step 4: VECTOR STORAGE
────────────────────────────
┌─────────────────────────────┐
│ FAISS Vector Store          │
│ • Index Type: Flat (L2)     │
│ • In-Memory Storage         │
│ • Per-Session Isolation     │
└──────┬──────────────────────┘
       │ Vector Store Ready
       ▼

Step 5: QUERY PROCESSING
────────────────────────────
┌──────────────┐
│ User Query   │
└──────┬───────┘
       │
       ▼
┌─────────────────────────────┐
│ Embed Query                 │
│ (same embedding model)      │
└──────┬──────────────────────┘
       │ Query Vector
       ▼

Step 6: SIMILARITY SEARCH
────────────────────────────
┌─────────────────────────────┐
│ FAISS Similarity Search     │
│ • Algorithm: Cosine Sim     │
│ • Retrieve Top K=3          │
│ • Return relevant chunks    │
└──────┬──────────────────────┘
       │ Relevant Context
       ▼

Step 7: PROMPT CONSTRUCTION
────────────────────────────
┌─────────────────────────────┐
│ LangChain Prompt Template   │
│ • System Instructions       │
│ • Context: Retrieved chunks │
│ • User Question             │
└──────┬──────────────────────┘
       │ Formatted Prompt
       ▼

Step 8: GENERATION
────────────────────────────
┌─────────────────────────────┐
│ Google Gemini 2.0 Flash     │
│ • Temperature: Default      │
│ • Max Tokens: Auto          │
│ • Response: Text            │
└──────┬──────────────────────┘
       │ Generated Answer
       ▼

Step 9: POST-PROCESSING
────────────────────────────
┌─────────────────────────────┐
│ • Save to chat history      │
│ • Format response           │
│ • Return to user            │
└──────┬──────────────────────┘
       │
       ▼
┌──────────────┐
│ User Sees    │
│ AI Response  │
└──────────────┘

Session Lifecycle Diagram

┌─────────────────────────────────────────────────────────────┐
│                    SESSION LIFECYCLE                         │
└─────────────────────────────────────────────────────────────┘

┌──────────────┐
│ User Opens   │
│ Application  │
└──────┬───────┘
       │
       ▼
┌─────────────────────────┐
│ Session Middleware      │
│ • Check session cookie  │
│ • No session exists?    │
└──────┬──────────────────┘
       │ YES - Create New
       ▼
┌─────────────────────────────────┐
│ sessionService.createSession()  │
│ • Generate unique ID (UUID)     │
│ • Initialize empty state:       │
│   - vectorStore: null           │
│   - chatHistory: []             │
│   - uploadedFiles: []           │
│   - createdAt: timestamp        │
│   - lastAccessedAt: timestamp   │
└──────┬──────────────────────────┘
       │ Session Created
       ▼
┌─────────────────────────────────┐
│ Store in Express Session        │
│ • Set cookie: connect.sid       │
│ • Save ragSessionId             │
└──────┬──────────────────────────┘
       │
       ▼
┌─────────────────────────────────┐
│ User Uploads Documents          │
│ • Files saved to uploads/       │
│ • Vector store created          │
│ • Linked to session             │
└──────┬──────────────────────────┘
       │
       ▼
┌─────────────────────────────────┐
│ User Chats (Multiple Times)     │
│ • Each message saved            │
│ • Chat history maintained       │
│ • lastAccessedAt updated        │
└──────┬──────────────────────────┘
       │
       ├─────────────┬──────────────────────┐
       │             │                      │
       ▼             ▼                      ▼
┌──────────────┐ ┌──────────────┐  ┌──────────────────┐
│ User Clears  │ │ User Closes  │  │ Session Timeout  │
│ Session      │ │ Browser      │  │ (30 min idle)    │
│ (Manual)     │ │ (Cookie exp) │  │ (Auto cleanup)   │
└──────┬───────┘ └──────┬───────┘  └──────┬───────────┘
       │                │                  │
       └────────────────┴──────────────────┘
                        │
                        ▼
       ┌────────────────────────────────┐
       │ sessionService.clearSession()  │
       │ • Delete uploaded files        │
       │ • Clear vector store (memory)  │
       │ • Clear chat history           │
       │ • Remove from sessions map     │
       └────────────────────────────────┘
                        │
                        ▼
       ┌────────────────────────────────┐
       │ Session Destroyed              │
       │ Resources freed                │
       └────────────────────────────────┘

Background Process:
┌─────────────────────────────────────────┐
│ Cleanup Service (runs every 10 min)    │
│ • Check all sessions                    │
│ • Find inactive > SESSION_TTL (30 min) │
│ • Auto-clear expired sessions           │
│ • Free memory resources                 │
└─────────────────────────────────────────┘

Deployment Architecture

┌─────────────────────────────────────────────────────────────┐
│                     DOCKER COMPOSE                          │
└─────────────────────────────────────────────────────────────┘

                    ┌──────────────────┐
                    │  Docker Network  │
                    │  rag-chatbot-net │
                    └────────┬─────────┘
                             │
            ┌────────────────┴────────────────┐
            │                                 │
            ▼                                 ▼
┌────────────────────────┐        ┌────────────────────────┐
│  Frontend Container    │        │  Backend Container     │
│  ────────────────────  │        │  ───────────────────   │
│  • Node:20-alpine      │        │  • Node:20-alpine      │
│  • Build: Vite         │        │  • Runtime: Node.js    │
│  • Serve: Nginx        │        │  • Express.js app      │
│  • Port: 80            │◄───────│  • Port: 5000          │
│  • Health: /health     │  API   │  • Health: /api/health │
└────────┬───────────────┘ Calls  └────────┬───────────────┘
         │                                  │
         │                                  │
    Serves Static                    ┌─────┴──────┐
    React App                        │            │
         │                           ▼            ▼
         │                  ┌─────────────┐ ┌─────────────┐
    User Browser            │ Volume:     │ │ Volume:     │
         │                  │ uploads/    │ │ cache/      │
         │                  │ (Persist)   │ │ (Models)    │
         └─────────────────►└─────────────┘ └─────────────┘
              Port 80
                                     │
                              External APIs
                                     │
                    ┌────────────────┴────────────────┐
                    │                                 │
                    ▼                                 ▼
           ┌──────────────────┐          ┌──────────────────┐
           │ Google Gemini    │          │ HuggingFace      │
           │ API (LLM)        │          │ Transformers     │
           │ gemini-2.0-flash │          │ all-MiniLM-L6-v2 │
           └──────────────────┘          └──────────────────┘

Technology Stack

Backend:

Express.js - Web framework
LangChain - RAG orchestration
FAISS - Vector store
HuggingFace Transformers - Text embeddings (all-MiniLM-L6-v2)
Google Gemini 2.0 Flash - LLM
Multer - File upload handling
Express Session - Session management

Frontend:

React 18 - UI framework
Vite - Build tool
Tailwind CSS - Styling
React Dropzone - File upload interface
React Markdown - Markdown rendering
Axios - HTTP client

Project Structure

rag-chatbot/
├── backend/                 # Express.js backend
│   ├── src/
│   │   ├── controllers/     # Request handlers
│   │   │   ├── uploadController.js
│   │   │   ├── chatController.js
│   │   │   └── sessionController.js
│   │   ├── services/        # Business logic
│   │   │   ├── ragService.js          # RAG implementation
│   │   │   ├── vectorStoreService.js  # FAISS operations
│   │   │   └── sessionService.js      # Session management
│   │   ├── routes/          # API routes
│   │   │   ├── uploadRoutes.js
│   │   │   ├── chatRoutes.js
│   │   │   └── sessionRoutes.js
│   │   ├── middleware/      # Express middleware
│   │   │   ├── uploadMiddleware.js    # Multer config
│   │   │   ├── sessionMiddleware.js
│   │   │   └── errorHandler.js
│   │   ├── utils/           # Utility functions
│   │   │   └── documentLoader.js      # PDF/DOCX loaders
│   │   └── app.js           # Express app entry point
│   ├── uploads/             # Uploaded files storage
│   ├── .env                 # Environment variables
│   ├── Dockerfile
│   └── package.json
│
├── frontend/                # React + Vite frontend
│   ├── src/
│   │   ├── components/      # React components
│   │   │   ├── ChatInterface.jsx
│   │   │   ├── FileUpload.jsx
│   │   │   ├── MessageInput.jsx
│   │   │   └── ErrorBoundary.jsx
│   │   ├── hooks/           # Custom hooks
│   │   │   ├── useChat.js
│   │   │   └── useSession.js
│   │   ├── services/        # API service
│   │   │   └── api.js
│   │   ├── App.jsx          # Main app component
│   │   └── main.jsx         # Entry point
│   ├── Dockerfile
│   ├── nginx.conf           # Nginx configuration
│   └── package.json
│
├── docker-compose.yml       # Docker orchestration
└── README.md                # This file

🚀 Quick Start with Docker (Recommended)

1. Clone and Setup

# Create .env file in root directory
cat > .env << EOF
GOOGLE_API_KEY=your_google_gemini_api_key_here
SESSION_SECRET=your_random_secret_string_here
EOF

2. Start with Docker Compose

docker-compose up -d

3. Access the Application

Frontend: http://localhost
Backend API: http://localhost:5000
Health Check: http://localhost:5000/api/health

4. Stop the Application

docker-compose down

🛠️ Local Development Setup

Backend Setup

Navigate to backend directory:

cd backend

Install dependencies:

npm install

Create .env file:

# Copy your Google Gemini API Key from: https://makersuite.google.com/app/apikey
GOOGLE_API_KEY=your_api_key_here

# Server Configuration
PORT=5000
NODE_ENV=development

# Session Configuration
SESSION_SECRET=your_random_secret_key_here

# Upload Limits
MAX_FILE_SIZE=10485760  # 10MB
MAX_FILES=5

# Session TTL (milliseconds) - 30 minutes
SESSION_TTL=1800000

Start the backend:

npm run dev

Backend will run on http://localhost:5000

Frontend Setup

Navigate to frontend directory:

cd frontend

Install dependencies:

npm install

Create .env file (optional):

VITE_API_URL=http://localhost:5000

Start the frontend:

npm run dev

Frontend will run on http://localhost:5173

📡 API Endpoints

Upload

POST /api/upload - Upload PDF/DOCX files
- Body: multipart/form-data with files field
- Response: Upload confirmation with file count and document count

Chat

POST /api/chat - Send a chat message
- Body: { "question": "Your question here" }
- Response: { "success": true, "data": { "answer": "AI response" } }
GET /api/chat/history - Get chat history
- Response: Array of chat messages with roles and timestamps

Session

GET /api/session - Get session information
- Response: Session details, uploaded files count, document count
DELETE /api/session - Clear current session
- Response: Confirmation and new session ID

Health

GET /api/health - Health check
- Response: Server status and active session count

🎯 Usage Guide

1. Upload Documents

Click or drag-and-drop PDF/DOCX files into the upload area
Wait for the success notification
Supported formats: .pdf, .docx
Limits: Max 5 files, 10MB each

2. Chat with Your Documents

Type your question in the input field at the bottom
Press Enter or click Send
AI will respond based on document content
Chat history is maintained during your session

3. Clear Session

Click the "Clear Session" button in the header to reset
This deletes all uploaded files and chat history
Session also clears automatically when you close the browser

⚙️ Configuration

Backend Environment Variables

Variable	Default	Description
`GOOGLE_API_KEY`	-	Required: Google Gemini API key
`PORT`	5000	Backend server port
`SESSION_SECRET`	-	Required: Secret for session encryption
`MAX_FILE_SIZE`	10485760	Max file size in bytes (10MB)
`MAX_FILES`	5	Max number of files per upload
`SESSION_TTL`	1800000	Session timeout in ms (30 minutes)
`NODE_ENV`	development	Environment mode
`FRONTEND_URL`	http://localhost:5173	Frontend URL for CORS

Frontend Environment Variables

Variable	Default	Description
`VITE_API_URL`	http://localhost:5000	Backend API URL

🔍 How It Works (RAG Pipeline)

1. Document Upload & Processing

User uploads PDF/DOCX → Backend receives files → 
Parse documents to text → Split into chunks (1000 chars, 200 overlap) →
Generate embeddings using HuggingFace (all-MiniLM-L6-v2) →
Store in FAISS vector store (in-memory, per session)

2. Question Answering

User asks question → Embed question using same model →
Search FAISS for top 3 most relevant chunks →
Combine chunks as context →
Send context + question to Google Gemini 2.0 Flash →
Generate answer based only on provided context →
Return answer to user

3. Session Management

Each user gets a unique session ID (stored in cookies)
Sessions store: vector store, uploaded files, chat history
Auto-cleanup runs every 10 minutes to remove inactive sessions
Sessions expire after 30 minutes of inactivity

🛠️ Troubleshooting

Common Issues

1. Upload fails with "Network Error"

Check if backend is running on port 5000
Verify CORS settings in backend/src/app.js
Ensure frontend is accessing correct API URL

2. "No documents uploaded" error when chatting

Upload documents first before asking questions
Check if upload was successful (green notification)
Try clearing session and re-uploading

3. Docker container fails to start

Verify .env file exists with required variables
Check if ports 80 and 5000 are available
Run docker-compose logs to see error details

4. Slow response times

First request is slow due to model loading
Subsequent requests should be faster
Consider increasing memory allocation for Docker

5. "Invalid file type" error

Only PDF and DOCX files are supported
Check file extension is exactly .pdf or .docx
Ensure file is not corrupted

📝 Development Notes

Adding New Document Types

To support additional file formats, modify:

backend/src/utils/documentLoader.js - Add new loader
backend/src/middleware/uploadMiddleware.js - Update file filter
frontend/src/components/FileUpload.jsx - Update accepted types

Changing the LLM

To use a different model, modify backend/src/services/ragService.js:

Update initializeGemini() method
Change model configuration
Adjust prompt template if needed

Customizing Embeddings

To use different embeddings, modify backend/src/services/vectorStoreService.js:

Change the modelName in getEmbeddings()
Options: any HuggingFace transformers model

🚀 Production Deployment

Docker Deployment (Recommended)

Set production environment variables
Build and run with Docker Compose:

docker-compose up -d --build

Use a reverse proxy (nginx/traefik) for HTTPS
Set NODE_ENV=production in backend
Configure proper session secrets

Manual Deployment

Backend:

cd backend
npm install --production
NODE_ENV=production node src/app.js

Frontend:

cd frontend
npm install
npm run build
# Serve dist/ folder with nginx or similar

📄 License

MIT License - feel free to use this project for personal or commercial purposes.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📧 Support

If you have any questions or run into issues:

Check the troubleshooting section above
Review the console logs for error details
Open an issue on GitHub

🔗 Links

Built with ❤️ using React, Express.js, LangChain, and Google Gemini

Session Management:
- Each user gets a unique session ID
- Vector store and chat history tied to session
- Inactive sessions cleaned up after 30 minutes
- Page refresh triggers session cleanup

🐛 Troubleshooting

Backend Issues

Error: "GOOGLE_API_KEY not found"

Ensure .env file exists in backend directory
Check that GOOGLE_API_KEY is set correctly

Error: "Failed to load documents"

Verify file format is PDF or DOCX
Check file size is under 10MB
Ensure backend has write permissions for uploads/ folder

Error: "Model download timeout"

First run downloads HuggingFace model (~50MB)
Wait a few minutes or check internet connection
Docker: Model cached in huggingface-cache volume

Frontend Issues

Files not uploading

Check backend is running on port 5000
Verify CORS settings in backend
Check browser console for errors

Chat not working

Ensure documents are uploaded first
Check backend logs for errors
Verify Gemini API key is valid

Docker Issues

Port already in use

Change ports in docker-compose.yml
Or stop conflicting services

Container fails to start

Check logs: docker-compose logs
Ensure .env file exists with API key
Try rebuilding: docker-compose up --build

📝 Development Notes

Vector Store: In-memory only, not persisted to disk
Embeddings: Cached in Docker volume for faster restarts
Session Storage: Express-session with memory store (use Redis for production scale)
File Cleanup: Uploaded files cleared with session (automatic)

🚀 Production Recommendations

Use Redis for session storage (scalability)
Implement rate limiting on API endpoints
Add authentication for multi-user support
Set up monitoring (health checks, logging)
Use HTTPS with valid SSL certificates
Configure nginx reverse proxy
Set resource limits in Docker (memory, CPU)
Implement file scanning for security

📄 License

MIT License - Feel free to use this project for personal or commercial purposes.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📧 Support

For issues and questions:

Create an issue on GitHub
Check existing issues for solutions

Built with ❤️ by Maitrek Patel

Name		Name	Last commit message	Last commit date
Latest commit History 9 Commits
backend		backend
frontend		frontend
.env.example		.env.example
.gitignore		.gitignore
README.md		README.md
docker-compose.yml		docker-compose.yml
setup.ps1		setup.ps1
setup.sh		setup.sh

maitrekpatel1612/docbot

Folders and files

Latest commit

History

Repository files navigation

RAG Chatbot - AI-Powered Document Assistant

🚀 Features

📋 Prerequisites

🏗️ Architecture

System Architecture Diagram

Data Flow Diagram

1. Document Upload Flow

2. Chat/Query Flow

Component Interaction Diagram

RAG Pipeline Diagram

Session Lifecycle Diagram

Deployment Architecture

Technology Stack

Project Structure

🚀 Quick Start with Docker (Recommended)

1. Clone and Setup

2. Start with Docker Compose

3. Access the Application

4. Stop the Application

🛠️ Local Development Setup

Backend Setup

Frontend Setup

📡 API Endpoints

Upload

Chat

Session

Health

🎯 Usage Guide

1. Upload Documents

2. Chat with Your Documents

3. Clear Session

⚙️ Configuration

Backend Environment Variables

Frontend Environment Variables

🔍 How It Works (RAG Pipeline)

1. Document Upload & Processing

2. Question Answering

3. Session Management

🛠️ Troubleshooting

Common Issues

📝 Development Notes

Adding New Document Types

Changing the LLM

Customizing Embeddings

🚀 Production Deployment

Docker Deployment (Recommended)

Manual Deployment

📄 License

🤝 Contributing

📧 Support

🔗 Links

🐛 Troubleshooting

Backend Issues

Frontend Issues

Docker Issues

📝 Development Notes

🚀 Production Recommendations

📄 License

🤝 Contributing

📧 Support

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Languages

Packages