Implement power spectrum density for filtering

[KedoKudo]

This PR introduces the following changes:

• Implemented power spectrum based hard thresholding and wiener filtering.
• Add new example notebooks to show case the usage.
• Simplify the top level API for regular users.
• Remove outdated functions.
• New modes available:
  • full: this is based on the 2020 paper which consists of six steps
  • simple: this is replacing the hard thresholding in full with a FFT based notch filtering for faster noise free pilot estimate.
  • express: this is based on the 2007 paper (two steps) and replacing the collaborative hard thresholding with a FFT based notch filter.

Testing instructions:

• Clone the repo and checkout the branch, make sure the testing data is also pulled via git submodule
• Run the three notebooks in the notebook folder
• Run the unit test with pytest -v at the root of the repo

EWM Story: 5228

[jmborr]

the tests and the python notebooks complete without error

[jmborr]

looks 👌