Local LLM Document Parser (Logistics Edition)

A privacy-focused, local document parser designed for the logistics/trucking industry. It uses Ollama (Llama 3.2) to intelligently categorize PDF documents into specific regulatory and operational folders without sending data to the cloud.

Features

• Local Intelligence: Uses llama3.2 running locally via Ollama. No API costs, no data privacy concerns.
• Smart Classification: Sorts documents into 12 specific categories:
  • IFTA
  • Corporation (Auto-detects California SOI)
  • IRP, Permits, Title Transfer, Driver Files, DOT
  • Clean Truck Check, Invoices, Drug Test
  • INFO (General)
• Manual Verification: Automatically routes ambiguous or low-confidence documents to a Manual_Verification folder for human review.
• Extensible: Built with a modular architecture to support future expansions like Excel parsing and Vector DB integration.

Prerequisites

• Python 3.10+
• Ollama: Must be installed and running.
  • Download Ollama or brew install ollama
  • Pull the model: ollama pull llama3.2

Installation

1. Clone the repository:

   git clone https://github.com/abhinyaay/document-parser.git
   cd document-parser

2. Create and activate a virtual environment:

   python3 -m venv venv
   source venv/bin/activate

3. Install dependencies:

   pip install -r requirements.txt

Usage

1. Start Ollama Service:

   brew services start ollama
   # OR simply run 'ollama serve' in a separate terminal

2. Add Documents: Place your PDF files in the input/ directory.

3. Run Parser:

   python3 parser_v2.py

4. Check Output: Sorted files will appear in the output/ directory, organized by category.

Project Structure

• parser_v2.py: Main entry point.
• classifier/local_llm.py: Logic interfacing with Ollama API.
• extractors/: Modules for reading different file formats (currently PDF).
• input/: Drop your raw files here.
• output/: Processed files land here.