Frequently asked questions

Citing asent

If you use this library in your research, please cite it using the following citation. Naturally you will have to change the version to fit the used version.

@software{Enevoldsen_Asent_Fast_flexible_2022,
author = {Enevoldsen, Kenneth},
title = {{Asent: Fast, flexible and transparent sentiment analysis}},
url = {https://github.com/KennethEnevoldsen/asent},
version = {0.4.0},
year = {2022}
}

or if you prefer APA:

Enevoldsen, K. (2022). Asent: Fast, flexible and transparent sentiment analysis (Version 0.4.0) [Computer software]. https://github.com/KennethEnevoldsen/asent

How do I test the code and run the test suite?

This package comes with an extensive test suite. In order to run the tests, you’ll usually want to clone the repository and build the package from the source. This will also install the required development dependencies and test utilities defined in the extras_require section of the pyproject.toml.

pip install -e ".[tests]"

python -m pytest

which will run all the test in the tests folder.

Specific tests can be run using:

python -m pytest tests/desired_test.py

If you want to check code coverage you can run the following:

python -m pytest --cov=.

How is the documentation generated?

Asent uses sphinx to generate documentation. It uses the Furo theme with custom styling.

To make the documentation you can run:

# install sphinx, themes and extensions
pip install -e ".[docs]"

# generate html from documentations

make -C docs html

How do I add a new language?

Check out the customizing Asent guide it gives an example of how to add a new language. Once you have added we would very much appreciate a pull request with your new language.

Couldn’t you just train a transformer for this?

It is true that a trained transformer would probably perform sentence-level sentiment analysis much better. However, this modular approach to sentiment analysis allows a more fine-grained sentiment analysis, as would for example be done in aspect-based sentiment analysis. This too, however, could be done using neural architectures. The goal of Asent it not to exclude such approaches (in fact it can use transformer-based models through spaCy) but to combine them in a way that allow the end user to examine the results.