Demo UC9.ipynb review

[soxofaan]

attempt to do review of Demo UC9.ipynb through a github issue:

• avoid doing star imports (it's practically always a bad idea):

  from openeo_classification.landuse_classification import *
  

• avoid long lines like these:

  train_test_split, algorithm, nrtrees, mtry, fusion_technique, aoi, strat_layer, start_date, end_date, nr_targets, nr_spp = getStartingWidgets()
  
  ml_model = X.fit_class_random_forest(target=json.loads(y.to_json()), training=train_test_split.value, num_trees=nrtrees.value, mtry=mtry.value)
  
  cube = features.filter_spatial(gpd.GeoDataFrame.from_features(json.loads(list(aoi.value.values())[0]["content"]))["geometry"][0])
  

  some notebook viewers (e.g. github) don't support horizontal scrolling, so you can't see the whole line

• functions (like the ones in openeo_classification.landuse_classification) should have a "snake_case" name instead of camel/dromedaris case:

  target_classes = getTargetClasses(nr_targets)
  ...
  y = getReferenceSet(aoi, nr_spp, target_classes)
  ...
  def getStrata():
  

  should be

  target_classes = get_target_classes(nr_targets)
  ...
  y = get_reference_set(aoi, nr_spp, target_classes)
  ...
  def get_strata():