snitc not breaking when wrong type of dataset is supplied

[lucas-johnson]

Looks like an awesome package that is going to be super helpful! I'm just getting started here. Thanks for building this!

Behavior

When I accidentally passed a <class 'numpy.ndarray'> instead of a <class 'rasterio.io.DatasetReader'> as the dataset argument to stmetrics.spatial.snitc() I get the following error:

Traceback (most recent call last):
  File "/Users/lucasj/Code/lib/stseg/test.py", line 14, in <module>
    segments = stmetrics.spatial.snitc(
  File "/Users/lucasj/Code/lib/stseg/venv/lib/python3.9/site-packages/stmetrics/spatial.py", line 100, in snitc
    for band in range(img.shape[0]):
UnboundLocalError: local variable 'img' referenced before assignment

Expected behavior

The program should break on this TypeError:

stmetrics/stmetrics/spatial.py

Line 96 in dd988d4

TypeError("Sorry we can't read this type of file. \

Suggestion

Consider using Raise to stop the execution like:

else:
    raise TypeError("Sorry we can't read this type of file. \
                     Please use Rasterio or xarray")

[andersonreisoares]

Thank you for bring this to our attention @lucas-johnson.

This will be fixed in our next commit. I'm also adding a test function to make sure this is working.