Project Mirage v0.1.2

Project Mirage is a command-line utility designed to organize and process photos and videos in a specified directory or directories. The project provides functionalities such as duplicate detection, culling, and organization.

Features

• Photos and Videos: Mirage can organize photos and videos. Works on multiple formats such as JPEG, PNG, HEIC/HEIF (iOS format), MP4, and MOV.

• Duplicate Grouping: Groups photos and videos based on specific criteria, enhancing the organization and accessibility of the files.

• Image Aesthetics Assessment: Using Vision Neural Models, Mirage can calculate which image from duplicates looks the best and cull them from the others.

Usage

Prerequisites

Software

• Python 3.x
• Required Python packages: torch, torchvision, Pillow, pillow_heif, pyiqa, mimetypes, argparse, uuid, tqdm
  • all packages can be installed with requirements.txt

Hardware

• At least 4GB of memory for vector embeddings
• Capable processor for vision models (CNNs like ResNet or VGG19)

I ran this on a Raspberry Pi 4 8GB with a 128GB SD card running Raspbian with Python 3.11.2 without any issues.

Installation

Clone the repository and install the required dependencies:

$ git clone https://github.com/hetkpatel/Project-Mirage.git
$ cd Project-Mirage
$ pip install -r requirements.txt

Run

$ python pipeline.py [dir] [--dry_run] [--show_unsupported_files] [--quality_check] [--order_by_date]

Arguments

• dir: One or more directory paths containing the images and videos to be processed. Mirage will crawl through all sub-directories from the parent directory and find all photos and videos to process.

• --dry_run (-d): Perform a dry run without any modifications. Useful for previewing changes.

• --show_unsupported_files (-u): Display unsupported files during execution.

• --quality_check (-q): Enable image quality checking.

• --order_by_date (-o): Enable image sort into sub-directories by date.

Example

$ python pipeline.py /path/to/images --dry_run --show_unsupported_files --quality_check --order_by_date

Output

Mirage generates an organized output structure in the current working directory:

• ./output/{session}/images: Processed and grouped images.

• ./output/{session}/videos: Processed and grouped videos.

CAUTION!!!

Use caution when running Mirage without the --dry_run option, as it will modify files and directories. Always review the dry run output before proceeding with actual modifications.

License

This project is licensed under the GNU AGPLv3 License - see the LICENSE file for details.