add bandwidth options

[jkingslake]

Irena's suggest was to have options for the start and stop frequency to use when computing the fft. This could look something like:

import xapres_package as xa
ds = xa.load.generate_xarray(directory='../../data/sample/single_dat_file/', 
             start_freq = [300e6, 400e6], 
             stop_freq = [500e6])

and this would do all the processing twice, once with start_freq = 200e6 and once with start_freq = 400e6. There would then be another dimension in the resulting xarray: start_freq.

If yo ualso provide multiple stop_freq's it could do every combination of start_freq and stop_freq.

cc @glugeorge

[jkingslake]

as a starting point though we could just allow start_freq and stop_freq to be adjustable parameters.

[jkingslake]

this should be possible using the new custom fft options that were added in #51 and a script I wrote for a different project (here), which allows you to loop over different parameter values and save the results in an xarray.

This wont result in a change to the code, I dont think, but I could add a notebook demonstrating it.

[jkingslake]

This works nicely.

You can download the script from github gist with a little package I found called getgist and then run the processing with

## load data
def load_zarr(f):
    return xr.open_dataset(f, engine = 'zarr', chunks = {})
LTG = load_zarr('/Users/jkingslake/Documents/science/ApRES/TG_apres/notebooks/zarrs/LTG_stacked.zarr').load()

## calculate velocities etc. 
dLTG = LTG.profile_stacked.displacement_timeseries().compute()

## get grid search function
!pip install getgist
!getgist jkingslake xarray_grid_search.py

## apply grid search function and plot
xgs.gridSearch(LTG.profile_stacked.displacement_timeseries, lower_limit_on_fit=[800, 1000, 1200, 1400]).strain_rate.plot(hue='lower_limit_on_fit')

cc @glugeorge, I wonder if this would be useful to you?

xgs.gridSearch(LTG.profile_stacked.displacement_timeseries, lower_limit_on_fit=[800, 1000, 1200, 1400])

outputs an xarray with a new dimension corresponding to the parameter you varied, in this case lower_limit_on_fit: