Intelligent Invoice Reimbursement System

An advanced system for automating invoice analysis and providing intelligent querying capabilities using Large Language Models (LLMs) and vector databases.

Features

• Automated Invoice Analysis: Process PDF invoices against company reimbursement policies
• Intelligent Data Storage: Store analysis results in a vector database for efficient retrieval
• Smart Querying: Natural language chatbot interface for querying invoice information
• Policy Compliance: Automated comparison of invoices against company policies

System Components

1. Invoice Analysis Endpoint

• Processes PDF invoices and company policies
• Uses LLM for intelligent analysis
• Determines reimbursement status (Fully/Partially Reimbursed or Declined)
• Stores results in vector database

2. RAG Chatbot Endpoint

• Natural language query interface
• Context-aware responses
• Intelligent retrieval of invoice information
• Support for metadata filtering

Installation

1. Clone the repository:

git clone https://github.com/Sibikrish3000/invoice-reimbursement-system.git
cd invoice-reimbursement-system

2. Create a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Set up environment variables: Create a .env file in the root directory with:

OPENAI_API_KEY=your_api_key_here
OPENAI_BASE_URL=your_endpoint_url

Usage

1. Start the server:

uvicorn app.main:app --reload

2. Access the API documentation at http://localhost:8000/docs

API Endpoints

Invoice Analysis

• POST /api/v1/analyze-invoice
  • Input: PDF files (policy and invoices), employee name
  • Output: Analysis results with reimbursement status

Chatbot Query

• POST /api/v1/chat
  • Input: Natural language query
  • Output: Markdown formatted response

Technical Details

Technologies Used

• Framework: FastAPI
• LLM Integration: LangChain with OpenAI
• Vector Store: ChromaDB
• Embedding Model: Sentence Transformers
• Document Processing: PyPDF

Architecture

• Modular design with separate services for invoice processing and querying
• Vector database for efficient similarity search
• RAG implementation for context-aware responses

Development

Project Structure

invoice-reimbursement-system/
├── app/
│   ├── main.py
│   ├── api/
│   │   ├── routes/
│   │   └── models/
│   ├── core/
│   │   ├── config.py
│   │   └── security.py
│   ├── services/
│   │   ├── invoice_processor.py
│   │   ├── vector_store.py
│   │   └── chatbot.py
│   └── utils/
│       ├── pdf_processor.py
│       └── embeddings.py
├── tests/
├── requirements.txt
└── README.md

License

Apache License 2.0

Containerization

To build and run your container:

docker build -t invoice-reimbursement-system .
docker run -p 8000:8000 --env-file .env invoice-reimbursement-system