HR Document Chat Bot

A Next.js application that allows users to upload HR documents and query them using natural language.

Important Production Notes

File Storage in Production

✅ Update: The application now uses Vercel Blob Storage for storing uploaded files. This makes it compatible with Vercel's read-only filesystem and provides a scalable solution for file storage.

Key changes:

1. Files are uploaded to Vercel Blob Storage
2. File URLs are stored in the PostgreSQL database
3. Document processing fetches files from Blob Storage when needed
4. Vector embeddings are stored in the database for efficient searching

⚠️ Important: You must set up Vercel Blob Storage and add the BLOB_READ_WRITE_TOKEN to your environment variables:

1. Create a Blob Store in your Vercel dashboard:

   • Go to your Vercel dashboard
   • Select your project
   • Go to Storage > Blob
   • Click "Create Blob Store"

2. Add the BLOB_READ_WRITE_TOKEN to your environment variables:

   BLOB_READ_WRITE_TOKEN="your-vercel-blob-read-write-token"
   

⚠️ Important: You must update your .env file to use a PostgreSQL database:

DATABASE_URL="postgresql://username:password@hostname:port/database"

If you prefer to use cloud storage instead of database storage for large files:

1. Sign up for a cloud storage service (AWS S3, Google Cloud Storage, etc.)
2. Update the file upload API route to upload files to the cloud storage
3. Update the file deletion logic to delete files from the cloud storage
4. Update the document processor to read files from the cloud storage

Features

• PDF Upload: Securely upload PDF documents through a user-friendly interface
• Document Processing: Extract and index text from PDF documents
• Natural Language Queries: Ask questions about your documents and get accurate answers
• Document Management: View and manage your uploaded documents
• User Authentication: Secure access with NextAuth.js

Technology Stack

• Frontend: Next.js, React, TailwindCSS
• PDF Processing: PDF.js
• Vector Database: PostgreSQL with vector storage for efficient similarity search
• AI: Google's Gemini model for generating responses and embeddings
• Authentication: NextAuth.js with Prisma adapter
• Database: PostgreSQL (both development and production)

Getting Started

Prerequisites

• Node.js 18.x or higher
• npm or yarn
• Google API key for Gemini

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/hr-bot.git
   cd hr-bot

2. Install dependencies:

   npm install

3. Set up environment variables:

   • Copy .env.example to .env.local and fill in the values

   cp .env.example .env.local

4. Initialize the database:

   npx prisma migrate dev

5. Start the development server:

   npm run dev

6. Open http://localhost:3000 in your browser to see the application.

Deploying to Vercel

Step 1: Prepare Your Database

The application uses SQLite by default, which is not suitable for production deployments on Vercel. For production, you need to use a PostgreSQL database:

1. Create a PostgreSQL database using Vercel Postgres, Supabase, Railway, or any other PostgreSQL provider.
2. Get your PostgreSQL connection string.

Step 2: Update Prisma Schema for Production

Before deploying, update your Prisma schema to use PostgreSQL:

1. Open prisma/schema.prisma
2. Change the database provider from sqlite to postgresql:

   datasource db {
     provider = "postgresql"
     url      = env("DATABASE_URL")
   }

3. Commit and push these changes to your repository.

Step 3: Deploy to Vercel

1. Push your code to GitHub
2. Create a new project in Vercel and link it to your GitHub repository
3. Set up the following environment variables in Vercel:
   • DATABASE_URL: Your PostgreSQL connection string
   • NEXTAUTH_SECRET: A secure random string for NextAuth.js
   • NEXTAUTH_URL: Your production URL (e.g., https://your-app.vercel.app)
   • GOOGLE_API_KEY: Your Google API key for Gemini AI
   • GEMINI_API_KEY: Your Gemini API key (if different from GOOGLE_API_KEY)
   • OAuth provider credentials if you're using social login
4. Deploy the application

Step 4: Run Migrations on Production Database

After deploying, you need to run migrations on your production database:

1. Install Vercel CLI:

   npm install -g vercel

2. Link your local project to the Vercel project:

   vercel link

3. Pull the environment variables:

   vercel env pull .env.production

4. Run the migrations:

   npx prisma migrate deploy

Usage

1. Upload Documents: Navigate to the Upload page to upload PDF documents.
2. Query Documents: Once documents are processed, go to the Query page to ask questions.
3. Manage Documents: View and manage your uploaded documents on the Documents page.

Security Considerations

• The application stores uploaded documents directly in the database as binary data.
• Document processing happens on the server, and the extracted text is stored in the database.
• Authentication is implemented using NextAuth.js.

Troubleshooting Vercel Deployment

If you encounter issues with Prisma during deployment:

1. Make sure the postinstall script is in your package.json:

   "postinstall": "prisma generate"

2. Make sure the build script includes Prisma generation:

   "build": "prisma generate && next build"

3. Check that your environment variables are correctly set in Vercel.

4. If you're seeing the error "Prisma has detected that this project was built on Vercel, which caches dependencies", make sure you have both the postinstall script and the updated build script.

5. If you're having issues with file uploads, remember that Vercel's filesystem is read-only in production. You'll need to use a cloud storage solution like AWS S3, Google Cloud Storage, or similar for storing uploaded files.

License

This project is licensed under the MIT License - see the LICENSE file for details.