Test calc rmax closes #48

[oceandie]

• Closes #xxxx
• Tests added
• Passes pre-commit run --all-files
• Changes are documented in whats-new.rst
• New functions/methods are listed in api.rst
• Project, label, and assignee tabs are populated

[oceandie]

https://www.dropbox.com/s/u3nl4htrynfi89x/bathymeter_mo_ps43.nc?dl=0
this is the link for AMM7 bathymetry ... let me know if you cannot use it

[oceandie]

I added the last implementation of xarray-calc_max (using the maximum rather than the average). I hope I am not doing anything stupidly wrong, if yes, sorry in advance :)

[jdha]

Not that I should be looking at this, but ...

To reproduce your SlopeParam code @oceandie [which looks very much like my old Matlab ;) ] I did the following (see below). The large values seem to appear in the outer halo as they where included in the np.diff (your code only goes from the second index to the last but one) ... I also think I'd got the indices out of sync by 1 - but I'm not too sure ... I'll look again on Monday (and do an XR version of this).

If I diff the the results from the following code with SlopeParam they match.

def calc_rmax_np(depth):
    """
    Calculate rmax: measure of steepness

    This function returns the slope steepness criteria rmax, which is simply
    (H[0] - H[1]) / (H[0] + H[1])
    Parameters
    ----------
    depth: float
            Bottom depth (units: m).
    Returns
    -------
    rmax: float
            Slope steepness value (units: None)
    """
    
    rmax_x, rmax_y = np.zeros_like(depth), np.zeros_like(depth)

    rmax_x[1:-1, 1:-1] = np.maximum (
        np.abs( np.diff(depth[1:-1, :-1], axis=1) / (depth[1:-1,  :-2] + depth[1:-1, 1:-1]) ),
        np.abs( np.diff(depth[1:-1,1:  ], axis=1) / (depth[1:-1, 1:-1] + depth[1:-1, 2:  ]) )
    )
    rmax_y[1:-1, 1:-1] = np.maximum (
        np.abs( np.diff(depth[ :-1, 1:-1], axis=0) / (depth[ :-2, 1:-1] + depth[1:-1, 1:-1]) ) ,
        np.abs( np.diff(depth[1:  , 1:-1], axis=0) / (depth[1:-1, 1:-1] + depth[2:,   1:-1]) )
    )

    return np.maximum(np.abs(rmax_x), np.abs(rmax_y))

[jdha]

for xr updated see #48