$\mu\text{TC}$
$\mu$TC is a minimalist text classifier, that is, it is designed to tackle text-classification problems in an agnostic way, being both domain and language independent.
$\mu$TC is intentionally simple, so only a small number of features were implemented. However, it uses a some complex tools from numpy and scikit-lean. The number of dependencies is limited and fulfilled by almost any Scientific Python distributions, e.g., Anaconda.
The documentation is on readthedocs.