This package contains a number of feature extraction methods on textual data, mostly expecting musical lyrics. As an work-in-progress development phase, we will continuously validate and update the package later on. Currently, several feature extraction methods including linguistic features
, topic_modeling
, psychology inventory based feature estimations
, etc.
To install this package, we recommend you use the python virtual environment. Inside the virtualenv, installation is using pip
and git
.
$ pip install git+https://github.com/mmc-tudelft/lyricpsych.git
feature extractor lyricpsych-extract
installed along with the package. The usage of the lyricpsych-extractor
can be found in the -h
option. For instance, you can extract personality
, value
, topic
, linguistic
features by using the example below:
$ lyricpsych-extract \
/path/to/the/lyrics_data.csv \
/path/for/output/ \
--w2v glove-twitter-25 \
--features linguistic value personality topic
Currently it's in its alpha version. It means some extractors are not fully validated, and may have a unexpected behavior. We will continue to work on improving those aspects, but also we are more than welcoming contributions. We are open to take issues and pull request.
- refactoring
- split extractor to dedicated extractors
- minor refactorings
- clean up
- unused functions
- unused data files
- unused scripts
- restructuring
- split
task
to the separate sub-module - separate
fm
andals_feat
to the separate repo
- split
- Documentation
- docstrings
- doc generation
- features
- experimental run reproduction cli
- deploy
- writing testings
- CI [Travis integration]
- register to PyPI
Jaehun Kim, Sandy Manolios
TBD
- Thanks to henryre, as our PLSA implementation of this API is extension of numba-plsa.