10 Quevedo Documentation

Quevedo is a python tool for creating, annotating and managing datasets of graphical languages, with a focus on the training and evaluation of machine learning algorithms for their recognition.

Quevedo is published under the Open Software License version 3.0, and the source code can be found online at:

• https://github.com/agarsev/quevedo

This is a reproduction of the documentation for Quevedo version 1.3 for the purposes of this thesis. The live version can be found here:

• https://agarsev.github.io/quevedo/latest/

Due to space concerns, the API for Quevedo as a library and the command line interface are not included here. They can be found online at https://agarsev.github.io/quevedo/latest/api/ and https://agarsev.github.io/quevedo/latest/cli/.

Installation can be performed using pip: pip install quevedo[web]. The web optional dependency is recommended, so as to be able to follow the discussion around visual annotation and the web interface.

Being a data-centric tool, having data on hand can make following this documentation easier. The VisSE corpus presented in this thesis can be used, and an example dataset can also be found along the source code, in the directory examples/toy_arithmetic of the repository.