Homogenization using kNN classifiers uses faulty weights in case SpectralHomogenizer.predict_by_machine_learner() is called with global_clf_threshold=None.

This is caused by https://gitext.gfz-potsdam.de/geomultisens/specclassify/-/issues/5 where nodata pixels are not flagged properly in this case.

Reproduce with:

SH = SpectralHomogenizer()
pred = SH.predict_by_machine_learner(
    arrcube=GeoArray('/any/L7_image.tif', nodata=-9999),
    method='LR',
    src_satellite='Landsat-7',
    src_sensor='ETM+',
    src_LBA=['1', '2', '3', '4', '5', '7'],
    tgt_satellite='Sentinel-2A',
    tgt_sensor='MSI',
    tgt_LBA=['1', '2', '3', '4', '5', '6', '7', '8', '8A', '11', '12'],
    classif_alg='kNN_MinDist',
    global_clf_threshold=None
)