Matic Lubej mlubej

## selection.py
def select_best_pair(self):
    """
    Select two individuals from the population randomly, where the probability of choosing an individual
    is weighted by it's fitness level.
    :return: two selected individuals
    """
    pop_fitness_list = np.array([ind.fitness for ind in self.population])
    pop_fitness_list = -pop_fitness_list + np.min(pop_fitness_list) + np.max(pop_fitness_list) + 1
    pop_fitness_list = pop_fitness_list / np.sum(pop_fitness_list)
    p1, p2 = np.random.choice(self.population,

## mutation.py
for idx, c in enumerate(chromosome):
    if np.random.rand() < mutation_probability:
        has_mutated = True
        rand_idx = np.random.choice(range(len(chromosome)))
        chromosome[idx], chromosome[rand_idx] = chromosome[rand_idx], chromosome[idx]

## get_unique_orientations.py
def get_unique_orientations(poly):
    """
    Out of all possible rotations and flips, take into account the polygon symmetry and
    obtain the unique polygon orientations.
    :param poly: input polygon
    :return: array of unique settings of (rotation, flip)
    """
    cx, cy = poly.centroid.x, poly.centroid.y
    pts = np.array(poly.checkers)[:, :-1]
    checkers = np.array(poly.checkers)[:, -1]

## evalscript.js
function setup() {
  return {
    input: ["sunZenithAngles", "sunAzimuthAngles"],
    mosaicking: Mosaicking.ORBIT,
    output: { id:"default", bands: 0}
  }
}

function updateOutputMetadata(scenes, inputMetadata, outputMetadata) {

## download_task_sun_angles.py
download_task = SentinelHubInputTask(
    bands_feature=(FeatureType.DATA, 'RGB'),
    bands = ['B04', 'B03', 'B02'],
    resolution=10,
    maxcc=1.0,
    time_difference=timedelta(hours=2),
    data_source=DataSource.SENTINEL2_L2A,
    max_threads=10,
    additional_data=[
        (FeatureType.MASK, 'dataMask'),

## eopatch_get_content.py
eopatch.timestamp
eopatch.mask['LULC']
eopatch.data['CLM'][0]
eopatch.data['BANDS'][5][..., [3, 2, 1]]

## eopatch_content.py
EOPatch(
  data: {
    BANDS: numpy.ndarray(shape=(80, 1010, 999, 6), dtype=float32)
  }
  mask: {
    CLM: numpy.ndarray(shape=(80, 1010, 999, 1), dtype=uint8)
    IS_DATA: numpy.ndarray(shape=(80, 1010, 999, 1), dtype=uint8)
    IS_VALID: numpy.ndarray(shape=(80, 1010, 999, 1), dtype=bool)
  }
  mask_timeless: {

## eopatch_load.py
# load EOPatch
EOPatch.load('path_to_eopatches/eopatch-0x6/')

## resampling_effecs.csv

          
            Model
            Overall Accuracy [%]
            F_1 Score [%]

            
              Uniform at 16-day rate
              94.4
              94.1

            
              Uniform at 8-day rate
              94.5
              94.3

            
              Optimally selected
              94.6
              94.4

## cloud_effects.csv

          
            Model
            Overall Accuracy [%]
            F_1 Score [%]

            
              No filtering/no masking
              92.8
              92.6

            
              No filtering/with masking
              94.2
              93.9

            
              With filtering/no masking
              94.0
              93.8

            
              With filtering/with masking
              94.4
              94.1
	def select_best_pair(self):
	"""
	Select two individuals from the population randomly, where the probability of choosing an individual
	is weighted by it's fitness level.
	:return: two selected individuals
	"""
	pop_fitness_list = np.array([ind.fitness for ind in self.population])
	pop_fitness_list = -pop_fitness_list + np.min(pop_fitness_list) + np.max(pop_fitness_list) + 1
	pop_fitness_list = pop_fitness_list / np.sum(pop_fitness_list)
	p1, p2 = np.random.choice(self.population,
	for idx, c in enumerate(chromosome):
	if np.random.rand() < mutation_probability:
	has_mutated = True
	rand_idx = np.random.choice(range(len(chromosome)))
	chromosome[idx], chromosome[rand_idx] = chromosome[rand_idx], chromosome[idx]
	def get_unique_orientations(poly):
	"""
	Out of all possible rotations and flips, take into account the polygon symmetry and
	obtain the unique polygon orientations.
	:param poly: input polygon
	:return: array of unique settings of (rotation, flip)
	"""
	cx, cy = poly.centroid.x, poly.centroid.y
	pts = np.array(poly.checkers)[:, :-1]
	checkers = np.array(poly.checkers)[:, -1]
	function setup() {
	return {
	input: ["sunZenithAngles", "sunAzimuthAngles"],
	mosaicking: Mosaicking.ORBIT,
	output: { id:"default", bands: 0}
	}
	}

	function updateOutputMetadata(scenes, inputMetadata, outputMetadata) {
	download_task = SentinelHubInputTask(
	bands_feature=(FeatureType.DATA, 'RGB'),
	bands = ['B04', 'B03', 'B02'],
	resolution=10,
	maxcc=1.0,
	time_difference=timedelta(hours=2),
	data_source=DataSource.SENTINEL2_L2A,
	max_threads=10,
	additional_data=[
	(FeatureType.MASK, 'dataMask'),
	eopatch.timestamp
	eopatch.mask['LULC']
	eopatch.data['CLM'][0]
	eopatch.data['BANDS'][5][..., [3, 2, 1]]
	EOPatch(
	data: {
	BANDS: numpy.ndarray(shape=(80, 1010, 999, 6), dtype=float32)
	}
	mask: {
	CLM: numpy.ndarray(shape=(80, 1010, 999, 1), dtype=uint8)
	IS_DATA: numpy.ndarray(shape=(80, 1010, 999, 1), dtype=uint8)
	IS_VALID: numpy.ndarray(shape=(80, 1010, 999, 1), dtype=bool)
	}
	mask_timeless: {
	# load EOPatch
	EOPatch.load('path_to_eopatches/eopatch-0x6/')
Model	Overall Accuracy [%]	F_1 Score [%]
Uniform at 16-day rate	94.4	94.1
Uniform at 8-day rate	94.5	94.3
Optimally selected	94.6	94.4