The Schelling Board Augmented Reality

This app helps play the Schelling Segregation Model Board Game by using augmented reality. It helps to play the game by using a phone or a tablet to check the happy and unhappy agents.

This app is part of the Transciències project. For information on how to get or play the board game, you can visit this page.

Launch app

You can launch the app using docker.

docker build -t schellingar_st_1_20_0 . 
docker run -d --restart unless-stopped -p 8501:8501 \ 
        -v ./data:/app/DataApp/data --name schellingar \
        schellingar_st_1_20_0 

where the data folder helps to keep the data persistent on the disk.

You can then access the app at http://localhost:8501

Configurations

To configure the app, you can use the config.ini and .env files located in the DataApp/ folder. The .env file contains database connection information, and the variable names begin with "SCHELLING_DB_". For instance, SCHELLING_DB_HOST refers to the database host. These variables take precedence over the values in the config.ini file and will override them.

Database

A database handles the matches and the data persistence. To initialize the schema you can run the following command

cd DataApp
python init_db.py

Train preparation

To train the model, we need first to create a dataset.

Cell division

We can take a picture of the board and divide the grid into cells.

For example, if we have a picture of a 21x25 grid, we can run

python Img2Game.py -i data/IMG_20221109_155318_499.jpg -o IMG_20221109_155318_499_jpg -t -g 21x25

where data/IMG_20221109_155318_499.jpg is the path to the image and 21x25 is the grid size. The output directory will be IMG_20221109_155318_499_jpg

Labelling

Then we can label each cell in the output directory in 5 different classes

• Team 1 - happy
• Team 2 - happy
• Team 1 - sad
• Team 2 - sad
• Empty cell

A simple technique for fast labelling is to move each cell in a directory with the name of the class.

Train/test split

Then, we need to split the picture into train and test sets. We can do this by running

splitfolders --output class_data_split --ratio .7 .2 .1 --seed 1234 -- class_data

from the library https://github.com/jfilter/split-folders

Contribute

Install

git clone git@github.com:dimitri-mar/SchellingBoardAR.git
cd SchellingBoardAR
pip install -r requirements.txt

Translation

Translations for both stramlit_app.py and streamlit_manager_app.py are based on gettext. Texts must be encased as _(string). Take into account that f-strings are not properly implemented. In case new texts in need of translation are added, the following steps must be taken:

• Create a .pot file with all strings for translation:

  Windows:

  python "path\pygettext.py" -d base -o DataApp\locales\base.pot DataApp\streamlit_manager_app.py DataApp\stramlit_app.py
  

  Unix:

  path/pygettext.py -d base -o DataApp/locales/base.pot DataApp/streamlit_manager_app.py DataApp/stramlit_app.py
  

• Update old sources with new strings

  Windows:

  path\msgmerge.exe locales\$language$\LC_MESSAGES\base.pot locales\base.pot -U
  

  Unix:

  msgmerge locales/$language$/LC_MESSAGES/base.pot locales/base.pot -U
  

• Add new translatations in base.pot file for each language.

• Create new .mo file for updated .pot file in each language

  Windows:

  "path\msgfmt.exe"-o DataApp\locales\$language$\LC_MESSAGES\base.mo DataApp\locales\$language$\LC_MESSAGES\base.pot
  

  Unix:

  msgfmt -o DataApp/locales/$language/LC_MESSAGES/base.mo DataApp/locales/$language/LC_MESSAGES/base.po
  

Acknowledgments

This app is part of the Transciències project by Heurística.

Please check the website for further information.

License

You can redistribute "The Schelling Board Augmented Reality" and/or modify it under the terms of the GNU Affero General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

All emojis are designed by OpenMoji – the open-source emoji and icon project. License: CC BY-SA 4.0