Pymodalib

This releases fixes some minor problems in the Python implementation of the wavelet transform.

This release reverts the change whereby the wavelet transform calculates `padleft` and `padright` in parallel; the RAM usage was often too high to justify the performance improvement.

Changes in this release:

- The windowed Fourier transform has been added.
- The wavelet transform now calculates `padleft` and `padright` in parallel, although this only provides a small performance improvement.
- The `harmonicfinder` algorithm is now more robust.
- PyMODAlib now checks that the MATLAB Runtime is installed, and the correct version, on Windows. This only occurs when a function which uses the MATLAB Runtime is called.

This release adds the Python implementation of the wavelet transform. Try running `examples/wavelet/wavelet-python.py` to test it out.

This release adds three new functions:

- `pymodalib.downsample` for downsampling a signal using a moving-average technique.
- `pymodalib.group_coherence` and `pymodalib.dual_group_coherence` for performing wavelet phase coherence on a group or pair of groups, with inter-subject surrogates.

This release adds:

- MATLAB implementation of the wavelet transform.
- Preprocessing algorithm for detrending signals.

Note: To use the wavelet transform, you need to install the MATLAB Runtime and PyMODA.