README.md

easy-vizy 😎

Minimal visualization tool to show images on a grid.

Useful for machine learning analysis as you can easily create interactive visualizations through URL linking.

Getting Started

Install flask

pip install flask

then, run the images server:

env FLASK_APP=main.py flask run -p 3000

If your images are not hosted, you can serve them locally using a simple python http server

cd images_dir #replace
python -m http.server

Take a look at main.py and update the image_id to URL mapping function if needed.

def get_image_urls(ids):
    """function to map ids to image URLS"""
    return [f"http://127.0.0.1:8000/{id}" for id in ids]

That's all!

Open http://localhost:3000/?asset_ids=image_id1,imageid2 ... with your browser to see the result.

Make sure to pass valid image ids from your dataset.

Using From a Notebook

You can load the grid visualization inline from your favorite notebook...

import IPython
# find out which images to visualize 
image_ids = ['image_id1', 'image_id2',...]
# construct the URL
url = "http://127.0.0.1:3000/?asset_ids=" + ",".join(image_ids)
# build an Iframe
iframe = '<iframe src='+url+', width=800, height=400></iframe>'

# Display the visualization 
IPython.display.HTML(iframe)

Customization

If you want to customize the visualizations, you can edit the index page (to add more info as needed) by modifying templates/index.html. Also, you can edit the sytles under static/style.css.

License

Easy-vizy is MIT licensed, as found in the LICENSE file. Please note that some of the dependencies AugLy uses may be licensed under different terms.