Fix zonal.stats() dask boolean short-circuit bug

[brendancol]

Bug

_stats_dask_numpy() in xrspatial/zonal.py (lines 243, 247) has a Python boolean short-circuit bug:

if 'mean' or 'std' or 'var' in stats_funcs:   # ALWAYS True
    compute_sum = True
    compute_count = True

if 'std' or 'var' in stats_funcs:              # ALWAYS True
    compute_sum_squares = True

In Python, 'mean' or 'std' or 'var' in stats_funcs evaluates as ('mean') or ('std') or ('var' in stats_funcs). The truthy string 'mean' short-circuits the entire expression, so compute_sum, compute_count, and compute_sum_squares are always set to True regardless of what statistics the user actually requested.

Impact

• Wasted computation: Users asking for just ['min', 'max'] on a large dask array pay for sum, count, and sum-of-squares aggregations they never asked for.
• Masked bugs: These code branches can never be disabled, so any future refactoring could introduce subtle data corruption without tests catching it.

Fix

if any(s in stats_funcs for s in ('mean', 'std', 'var')):
    compute_sum = True
    compute_count = True

if any(s in stats_funcs for s in ('std', 'var')):
    compute_sum_squares = True

Add tests that verify only the requested statistics trigger computation.