CTA Orange

CTA Orange is an add-on for Orange Canvas created for Computational Text Analysis and based on CTA Kernel. It adds 8 new widgets meant to propose an interface based edit of a CTA Kernel's workflow. It is designed to illustrate the possibility of making scientific claims about text analysis in an interactive, visual, and executable way. The bundled example uses synthetic data to illustrate a workflow comparing the diversity of emoji strings of different lengths and evaluating whether the resulting evidence supports a claim.

Requirements

• Python ≥ 3.12
• Orange3 3.40.0
• CTA Kernel 0.1.0

For development from source, uv is also required.

Installation

For regular use with the 0.1.1 release, install the package from PyPI:

python -m pip install cta-orange==0.1.1

For development from a repository checkout, open a terminal in the folder containing pyproject.toml and run:

uv sync

Running

After a regular installation, launch Orange with:

python -m Orange.canvas

From a development checkout managed with uv, use:

uv run python -m Orange.canvas

Documentation

Each widget has its own documentation under docs/widgets.

Development context

These widgets have been created by Colin Luginbühl in the context of a UNIL course under the supervision of Aris Xanthos who designed the CTA Kernel infrastructure.
AI assistants, including Claude, Deepseek, and ChatGPT models, have been used for reviewing parts of the code and documentation. Nevertheless, the authors declare full responsibility for the created widgets and documentation.