LlavaNetScanner

Project Overview

LlavaNetScanner is an AI-enabled swarm that utilizes multiple agents to analyze various design files (e.g., images, PDFs, DWG/DXF files). The project automates file scanning, generates detailed descriptions based on file content and metadata, and organizes the results into a database and CSV files. A web-based interface, built with Streamlit, allows you to view, filter, and manage the scanned data interactively.

Key Features

1. Automated Scanning: The system automatically scans directories for supported design files.
2. File Analysis & Description:
   • Images: Analyzed via Llava to generate detailed, concise, or creative descriptions.
   • PDFs: Processed by Llama3 (when text is extractable) or converted and analyzed via Llava.
   • CAD Files: DWG/DXF file names are parsed for basic description, with plans for future in-depth analysis.
3. Metadata Management: Extracts and manages metadata including file paths, contractors, projects, file sizes, and timestamps.
4. Database Integration: Stores scan results in an SQLite database and includes utilities for migrating data to CSV or PostgreSQL.
5. Interactive Streamlit Interface: View database contents, apply filters, visualize data distributions, and initiate new directory scans via an intuitive web UI.
6. Additional Utilities:
   • CSV to PostgreSQL Migration: Migrate CSV data to a PostgreSQL database.
   • File Storage Analysis: Use the netgrapher tool to analyze storage usage by file type.

Components

• app.py: The main Streamlit application that provides an interactive UI to view the database, visualize statistics, and scan new directories.
• NetScanner.py: Handles file scanning and processing—delegating tasks based on file type.
• image_describer.py: Generates image descriptions using Llava.
• design_describer.py: Provides basic descriptions for CAD files (.dwg, .dxf) based on file name cues.
• pdf_describer.py: Processes and describes PDF files.
• database.py: Manages the SQLite database that stores scan results.
• csv_to_postgres.py: Script to migrate CSV data into a PostgreSQL database.
• migrate_csv_to_db.py: Utility to migrate existing CSV data into the local database.
• netgrapher.py: Analyzes a directory's file distribution and visualizes storage usage with a pie chart.
• LICENSE: GNU General Public License.

Installation

1. Clone the Repository:

   git clone <repository_url>
   cd LlavaNetScanner

2. Install Dependencies:
   Ensure you have Python 3.7+ installed, then run:

   pip install -r requirements.txt

   If a requirements.txt file is not provided, manually install the required packages:

   pip install streamlit pandas plotly psycopg2 matplotlib numpy python-dotenv ollama

3. Set Up Models:

   • Install Ollama: pip install ollama
   • Ensure Llava and Llama3 models are available by running:

     Ollama run Llava
     Ollama run Llama3

Usage

Running the Interactive Web Interface

The primary interface is built with Streamlit. To start the application, run:

```bash
 streamlit run app.py --server.port 8502 --server.address localhost
 ```