Naplib

-------------

- |Fix| : The ``auditory_spectrogram`` function was giving incorrect results when the input sampling rate was not a multiple of 2. For consistency with the Matlab wav2aud function's output, the function now resamples audio to 16k sampling rate before computing the auditory spectrogram, which ensures that the output is in the correct frequency range.
- |Fix| : Multiple miscellaneous fixes for the ``process_ieeg`` pipeline edge cases.

-------------

- |Efficiency| : The major functionality of ``filterbank_hilbert`` has been significantly optimized and put into a new function called ``filter_hilbert`` which reduces memory usage by averaging the output over center frequencies before returning it to the user, utilizing up to 50x less memory.
- |API| : The API of filtering functions within the ``preprocessing`` module has changed to support inplace operations or other API changes that may change their output compared to v0.1.10 by changing default values of arguments. This includes ``filter_line_noise``
- |MajorFeature| : New module called ``naplab`` containing preprocessing pipelines and tools used by the Neural Acoustic Processing Lab (NAPLab) for processing raw neural data.

--------------

- |Fix| : Changed the ``features.get_wrd_dict`` function to create a dictionary which does not use the value 0 for any words, which fixes an issue when performing word alignment where one word in the transcript might be assigned the value of 0, which would be masked by the 0's which indicate spacing.

-------------

- |Feature| : Added the ``kdeplot`` function to ``naplib.visualization`` which plots kernel density and histograms jointly, and for multiple distributions at once.
- |Enhancement| : Expanded the functionality of ``naplib.visualization.shadederrorplot`` to allow computing the confidence interval using percentiles (such as 95% confidence interval), and to allow plotting the median or the mean at each time point.
- |API| : All visualization functions (except the default case of ``hierarchicalclusterplot`` given its multi-axis nature) now return the axes on which the data were plotted.

-------------

- |Feature| : Added the ability to read an HTK file with ``naplib.io.read_htk``.
- |Enhancement| : Expanded the English phonetic dictionary file used by ``features.Aligner`` to include more words.
- |Fix| : Creating an empty ``naplib.Data`` object by initializing it with no arguments no longer raises an exception. This allows you to create an empty Data object and then build it up easily from a blank starting point.

-------------

- |Fix| : Fix issue where ``stats.responsive_ttest`` to allow customization of the time periods to compare between before and after stimulus onset to test for stimulus responsiveness. Also fix a minor issue where p-values where not properly corrected for multiple tests.