Smart Invoice Matching

Contributors

• Anup Nalwade
• Vitthal Biradar
• Bharat Patil
• Sumit Tippanbone

---

Smart Invoice Matching is a Flask-based web service that processes purchase orders and invoices. It leverages machine learning techniques—including TensorFlow and scikit-learn—to perform matching and generate reports. The application also uses Cloudinary for file storage and MongoDB for record keeping.

Features

• File Processing: Accepts CSV or Excel files for purchase orders and invoices.
• Data Matching: Uses TF-IDF vectorization, cosine similarity, and additional heuristics (SequenceMatcher) to match invoice records.
• Machine Learning: Trains a neural network model (using TensorFlow) to evaluate and classify matches.
• Dimensionality Reduction: Implements dynamic dimensionality reduction with TruncatedSVD.
• Report Generation: Generates matching reports in Excel format and uploads them to Cloudinary.
• API Endpoints:
  • /process: Processes PO and Invoice files, performs matching, trains the model, and stores report data.
  • /manual_match: Allows manual match corrections.
  • /fetch_entries: Retrieves stored matching entries from MongoDB.
  • /process_xlsx: Processes and filters Excel reports with pagination.

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/smart-invoice-matching.git
   cd smart-invoice-matching