I built this generic SVM solver for my uses in kaggle. It supports entering any dataset and the SVM attempts to learn on it, provided that it is in the correct format.
If the format is not proper in the raw dataset, it allows the user to import any dataset after it has been put through the users custom modeler.
See the example/ folder for an example on using the SVM class as well as a very basic modeler.
All files are documenting using docstring. Any merge requests should also be formatted with this.