🧠 Smart Invoice Classifier (Python + Machine Learning)

A machine learning–based system that automatically extracts text from invoices, processes it using NLP, and classifies invoices into categories such as Rent, Utilities, Supplies, etc.
It provides a REST API for uploading invoices and retrieving classified results, and securely stores all data in a database for analysis.

---

📌 Project Overview

This project demonstrates how to build an end-to-end invoice classification system using:

• OCR (Tesseract) to extract text from PDFs
• NLP + Machine Learning to classify text into categories
• Flask REST API to expose endpoints for uploading and retrieving invoices
• MongoDB to store extracted text and classification results
• Secure data handling (temporary storage, masking sensitive fields)

---

📂 Project Structure

smart-invoice-classifier/
│
├── data/
│   ├── dummy_invoices/          # Generated sample invoices
│   └── labels.csv               # Ground truth labels for training
│
├── src/
│   ├── extract/
│   │   └── pdf_extractor.py     # OCR and text extraction scripts
│   ├── model/
│   │   └── classifier.py        # Training and inference models
│   ├── api/
│   │   └── routes.py            # Flask API endpoints
│   ├── utils/
│   │   └── __init__.py          # Helper functions
│   └── database/
│       └── mongo_db.py          # DB configuration and operations
│
├── tmp_uploads/                 # Temporary storage for uploaded files
├── requirements.txt
├── README.md
└── main.py                      # Main application entry point

---

🔧 Key Components

1. Data Extraction (src/extract/pdf_extractor.py)

• Uses Tesseract OCR for text extraction from PDF invoices
• Handles various invoice formats and layouts
• Preprocesses images for better OCR accuracy

2. Machine Learning Model (src/model/classifier.py)

• NLP preprocessing (tokenization, TF-IDF, word embeddings)
• Multi-class classification (SVM, Random Forest, or Neural Networks)
• Model training and evaluation pipelines

3. REST API (src/api/routes.py)

• POST /upload - Upload invoice PDFs

• GET /classified - Display invoices from database

4. Database (src/database/mongo_db.py)

• MongoDB integration for storing invoice data
• Secure storage of extracted text and classifications
• Efficient querying and retrieval operations

---

Demo Video

demo.video.mp4

Commit History

Note

• The project configuration (e.g., port number, MongoDB URL, etc.) is hardcoded instead of using an .env file to allow the recruiter to run the code directly without additional setup.