Make sure the Google Cloud SDK is installed and configured.

Install requirements:

pip install flask gunicorn tqdm

First, collect raw sizes for every object in every bucket. The following script does that and stores the results to the data/ directory.

bash collect_usage.sh

Now, run the Flask app, which reads from data/, builds a tree data structure in memory, and serves a simple interactive explorer of that tree. For debugging, you can use python app.py. For slightly more productionized serving, I use gunicorn with the --preload option, which shares the data structure between workers to save memory.

gunicorn "app:app" --preload -w <num_workers>