Cullify

Cullify is a photo culling tool focused on speed, trust, and a clean workflow. It detects unreadable files, blur, near-duplicates, and closed eyes.

Current repo layout (simplified)

• app.py — Streamlit prototype UI
• cullify.py — CLI runner
• filters/ — blur, duplicates, eyes detection
• models/ — face cascade (haarcascade_frontalface_default.xml)
• bad/, filtered/, runs/ — local output/test folders (ignored)
• .gitignore — keeps local data and caches out of git

Quick start

1. Create and activate a virtual environment

• macOS/Linux:
  • python3 -m venv .venv
  • source .venv/bin/activate
• Windows (PowerShell):
  • py -m venv .venv
  • .venv\Scripts\Activate.ps1

2. Install dependencies

• pip install -r requirements.txt

3. Run the Streamlit app (prototype)

• streamlit run app.py

4. Or run the CLI

• python cullify.py --input /path/to/images --output /path/to/output

Notes

• The filters use OpenCV, Pillow, and MediaPipe (for eyes). Make sure your environment has these installed (see requirements.txt).
• The models/ folder should contain haarcascade_frontalface_default.xml.
• Local data folders like runs/, bad/, filtered/ are ignored by git to keep the repo clean.