List view

• 0.3

  Features projected to wind up in the next major release (0.3).

  No due date

  •13/23 issues closed
  56% complete10 open13 closed

• 0.2

  This will be the next major release. Target features to add include: * A `RemoteStim` hierarchy or some such, that allows us to represent remote resources that can be retrieved. * Support for integrated retrieval of data from major social media APIs (e.g., Twitter and Facebook), ideally with associated `Stim` classes. * Scikit-learn integration: `Graph` instances should be injectable into existing scikit-learn pipelines (e.g., by wrapping them with a class that supports the sklearn `TransformerMixin` pattern). * Basic parallelization (a very crude implementation is already included, via #106), but we could do a better job parallelizing branches in `Graph`s, testing more comprehensively, etc. * More systematic documentation for `Extractor` classes: ideally, we should settle on some standard way to represent all key values returned by an `Extractor` inside its docstring. * Support for several other feature extraction toolboxes implemented in Python, with particular attention to music-related features, as there are a ton of tools out there. * Support for a few more APIs from the major players—e.g., Amazon's Rekognition service. * Generalized support for the TensorFlow model zoo (we currently support only one pretrained model), and possibly other Python-based model zoos. * More comprehensive documentation generated via Sphinx and/or hosted on readthedocs.

  No due date

  •11/11 issues closed
  100% complete0 open11 closed