Sliding window changes

[mibur1]

Hi @mtorabi59

I implemented the changes as discussed in #31.

The changes are:

• Removed extending the window which has two effects:
  • Bugfix regarding the zero-padding for rectangular SW calculation
  • Limits the tapered window to contain no samples outside the window width
• Added a window_std parameter to Sliding Window, Discrete HMM, and Sliding Window Clustering
  • Default is None, which results to the original std of 3 * W / 22
• Disabled the convolution and instead directly use the Gaussian
  • In my opinion this is unnecessary now, as the window size is a hard limit for the window
• Vectorized Pearson correlation calculation for improved performance

Feel free to modify this to your liking. Also I did not write any tests/documentation, please let me know if I should do so :)

[mtorabi59]

Thanks @mibur1 ! Overall it looks good! Thank you for making the changes. There a few comments below:

[mtorabi59]

@mibur1 lmk if you have any Qs!
Also, tests are not necessary, but if you have some ideas for testing the sliding window, it'd be great to have them!

[mibur1]

@mtorabi59 Thanks for the quick review and comments, I hope I addressed all of them. I will think about tests for the sliding window and add some if I have an idea.

[mtorabi59]

@mibur1 Thanks, everything looks good! There might only be some formatting changes by black or flake8.

[mtorabi59]

@mibur1 If you're familiar with pre-commit, it would be great to run it locally and make the required formatting changes, ow I'll fix them!

[mibur1]

Thanks! I'll try to do this as well as the unit tests next week. But please feel free to do it yourself if you would like to have it done earlier.

[mtorabi59]

@mibur1 Thanks !

[mtorabi59]

@mibur1 Since the remaining changes are very small and you have done most of the work, I'm just gonna make the final adjustments myself.
Thanks again for your contribution!

[mibur1]

Sorry, I forgot about that. Thanks a lot for handling it!