Sustainable Fashion Assistant

Ecostyle_demo.mov

Step 1: Clone the repository

• git clone https://github.com/jitani04/ecostyle.git

Step 2: Navigate to the project directory

• cd ecostyle

Step 3: Install the necessary dependencies

• npm install

Step 4: Ensure you have the proper .env file

To connect to the Supabase, it should have the following:

• VITE_SUPABASE_PROJECT_ID
• VITE_SUPABASE_PUBLISHABLE_KEY
• VITE_SUPABASE_URL
• SUPABASE_SECRET_KEY
• VITE_RECOMMEND_API_URL="http://localhost:8000/search"

Note: to access our database, please reach out and we can provide the keys

Step 4: Start the development server

The following commands use auto-reloading and provide an instant preview:

• cd frontend
• npm run build

After the build completes, the output will be in frontend/dist/ In Chrome/Edge/Brave open Extensions -> Load unpacked and select the frontend/dist/ folder.

Step 4: Run the backend server

• cd backend
• create a python venv
  • python -m venv eco-venv
  • source eco-venv/bin/activate
• install dependencies from requirements.txt
  • pip install -r requirements.txt
• if you cant run the requirements.txt file, run these commands manually:

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
pip install Pillow requests python-dotenv supabase ftfy regex timm
pip install git+https://github.com/openai/CLIP.git
pip install fastapi uvicorn
pip install python-multipart
pip install numpy==1.26.4

use these commands to run the backend:

• cd clothing-rec-program
• uvicorn CLIP_text_rec_backend:app --reload

go to the link to access the backend and test it out, click on the /search endpoint and upload an image of a product image:

• http://localhost:8000/docs

you can now go to your favorite online clothing website and test the extension out!

This project is built with:

• Vite
• TypeScript
• React
• FastAPI
• Python
• shadcn-ui
• Tailwind CSS

Final Presentation:

Ecostyle Final Poster.pdf