paprika

Paprika is a project for the Museum of Tübingen, where a team of students visualize how a neural image classifer works.

It was on display from 02/2023 - 21/01/2024 (https://www.tuebingen.de/stadtmuseum/38998.html).

Installation instructions

First download poetry. Then, with poetry install, switch to the directory root. Activate a shell using

poetry shell

Then, install paprika using

poetry install

If you use VSCode, to get the venv to be picked up, type

poetry config virtualenvs.in-project true

Before creating the venv. If you have already created the venv, delete it:

poetry env remove <current environment>

And then install the venv again.

To run the UI, run the run-ui.py in the scripts folder. You might have to change the configuration under paprika/ui/_config.py if your setup differs from the one used in the museum.

Repository structure

• scripts contains everything that is directly executable and serves some testing purpose
• paprika is the main module and contains three submodules: ai, cam and ui
• In each submodule, module's exports gets explicitly reimported in init.py. Module-internal files get prefaced with an underscore

Contributor Reminders

• Format all code with black before commiting
• Use types in Python code where possible, check typing errors with pyright
• Add your dependencies to poetry, using poetry add or by hand
• Don't drop files into the repository root, use appropriate sub folders
• Write basic documentation where possible
• Write tests (under a test folder) if possible
• If you change code, see if the tests still run
• Write sensible commit messages that precisely explain your changes
• Write commits that serve one purpose (f.e. don't mix up formatting/refactoring and adding features)
• Develop larger features on a separate branch. Always leave the main branch in an intact state