def test_multi_processing(self):
task = SentinelHubInputTask(
bands_feature=(FeatureType.DATA, 'BANDS'),
bands=["B01", "B02", "B03"],
additional_data=[(FeatureType.MASK, 'dataMask')],
size=self.size,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_collection=DataCollection.SENTINEL2_L1C,
max_threads=self.max_threads
)
time_intervals = [
('2017-01-01', '2017-01-30'),
('2017-02-01', '2017-02-28'),
('2017-03-01', '2017-03-30'),
('2017-04-01', '2017-04-30'),
('2017-05-01', '2017-05-30'),
('2017-06-01', '2017-06-30')
]
with futures.ProcessPoolExecutor(max_workers=3) as executor:
tasks = [executor.submit(task.execute, None, self.bbox, interval) for interval in time_intervals]
eopatches = [task.result() for task in futures.as_completed(tasks)]
array = np.concatenate([eop.data['BANDS'] for eop in eopatches], axis=0)
width, height = self.size
self.assertTrue(array.shape == (20, height, width, 3))
def test_additional_data(self):
""" Download additional data, such as viewAzimuthMean, sunAzimuthAngles...
"""
task = SentinelHubInputTask(bands_feature=(FeatureType.DATA, 'BANDS'),
bands=['B01', 'B02', 'B05'],
additional_data=[
(FeatureType.MASK, 'dataMask',
'IS_DATA'), (FeatureType.MASK, 'CLM'),
(FeatureType.MASK, 'SCL'),
(FeatureType.MASK, 'SNW'),
(FeatureType.MASK, 'CLD'),
(FeatureType.DATA, 'CLP'),
(FeatureType.DATA, 'viewAzimuthMean',
'view_azimuth_mean'),
(FeatureType.DATA, 'sunAzimuthAngles'),
(FeatureType.DATA, 'sunZenithAngles')
],
size=self.size,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_source=DataSource.SENTINEL2_L2A,
max_threads=self.max_threads)
eopatch = task.execute(bbox=self.bbox,
time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, 'BANDS')]
is_data = eopatch[(FeatureType.MASK, 'IS_DATA')]
clm = eopatch[(FeatureType.MASK, 'CLM')]
scl = eopatch[(FeatureType.MASK, 'SCL')]
snw = eopatch[(FeatureType.MASK, 'SNW')]
cld = eopatch[(FeatureType.MASK, 'CLD')]
clp = eopatch[(FeatureType.DATA, 'CLP')]
view_azimuth_mean = eopatch[(FeatureType.DATA, 'view_azimuth_mean')]
sun_azimuth_angles = eopatch[(FeatureType.DATA, 'sunAzimuthAngles')]
sun_zenith_angles = eopatch[(FeatureType.DATA, 'sunZenithAngles')]
self.assertTrue(
np.allclose(array_stats(bands), [0.027, 0.0243, 0.0162]))
width, height = self.size
self.assertTrue(bands.shape == (4, height, width, 3))
self.assertTrue(is_data.shape == (4, height, width, 1))
self.assertTrue(is_data.dtype == np.bool)
self.assertTrue(clm.shape == (4, height, width, 1))
self.assertTrue(clm.dtype == np.uint8)
self.assertTrue(scl.shape == (4, height, width, 1))
self.assertTrue(snw.shape == (4, height, width, 1))
self.assertTrue(cld.shape == (4, height, width, 1))
self.assertTrue(clp.shape == (4, height, width, 1))
self.assertTrue(view_azimuth_mean.shape == (4, height, width, 1))
self.assertTrue(sun_azimuth_angles.shape == (4, height, width, 1))
self.assertTrue(sun_zenith_angles.shape == (4, height, width, 1))
self.assertTrue(len(eopatch.timestamp) == 4)
def test_additional_data(self):
task = SentinelHubInputTask(
bands_feature=(FeatureType.DATA, "BANDS"),
bands=["B01", "B02", "B05"],
additional_data=[
(FeatureType.MASK, "dataMask", "IS_DATA"),
(FeatureType.MASK, "CLM"),
(FeatureType.MASK, "SCL"),
(FeatureType.MASK, "SNW"),
(FeatureType.MASK, "CLD"),
(FeatureType.DATA, "CLP"),
(FeatureType.DATA, "viewAzimuthMean", "view_azimuth_mean"),
(FeatureType.DATA, "sunAzimuthAngles"),
(FeatureType.DATA, "sunZenithAngles"),
],
size=self.size,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_collection=DataCollection.SENTINEL2_L2A,
max_threads=self.max_threads,
)
eopatch = task.execute(bbox=self.bbox,
time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, "BANDS")]
is_data = eopatch[(FeatureType.MASK, "IS_DATA")]
clm = eopatch[(FeatureType.MASK, "CLM")]
scl = eopatch[(FeatureType.MASK, "SCL")]
snw = eopatch[(FeatureType.MASK, "SNW")]
cld = eopatch[(FeatureType.MASK, "CLD")]
clp = eopatch[(FeatureType.DATA, "CLP")]
view_azimuth_mean = eopatch[(FeatureType.DATA, "view_azimuth_mean")]
sun_azimuth_angles = eopatch[(FeatureType.DATA, "sunAzimuthAngles")]
sun_zenith_angles = eopatch[(FeatureType.DATA, "sunZenithAngles")]
assert calculate_stats(bands) == approx([0.027, 0.0243, 0.0162])
width, height = self.size
assert bands.shape == (4, height, width, 3)
assert is_data.shape == (4, height, width, 1)
assert is_data.dtype == bool
assert clm.shape == (4, height, width, 1)
assert clm.dtype == np.uint8
assert scl.shape == (4, height, width, 1)
assert snw.shape == (4, height, width, 1)
assert cld.shape == (4, height, width, 1)
assert clp.shape == (4, height, width, 1)
assert view_azimuth_mean.shape == (4, height, width, 1)
assert sun_azimuth_angles.shape == (4, height, width, 1)
assert sun_zenith_angles.shape == (4, height, width, 1)
assert len(eopatch.timestamp) == 4
def test_no_data_input_task_request(self):
task = SentinelHubInputTask(
bands_feature=(FeatureType.DATA, "BANDS"),
additional_data=[(FeatureType.MASK, "dataMask")],
size=self.size,
maxcc=0.0,
data_collection=DataCollection.SENTINEL2_L1C,
)
eopatch = task.execute(bbox=self.bbox,
time_interval=("2021-01-01", "2021-01-20"))
bands = eopatch[FeatureType.DATA, "BANDS"]
assert bands.shape == (0, 101, 99, 13)
masks = eopatch[FeatureType.MASK, "dataMask"]
assert masks.shape == (0, 101, 99, 1)
def test_S2L1C_float32_uint16(self):
""" Download S2L1C bands and dataMask
"""
test_dir = os.path.dirname(os.path.realpath(__file__))
cache_folder = os.path.join(test_dir, 'cache_test')
if os.path.exists(cache_folder):
shutil.rmtree(cache_folder)
task = SentinelHubInputTask(
bands_feature=(FeatureType.DATA, 'BANDS'),
additional_data=[(FeatureType.MASK, 'dataMask')],
size=self.size,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_collection=DataCollection.SENTINEL2_L1C,
max_threads=self.max_threads,
cache_folder=cache_folder)
eopatch = task.execute(bbox=self.bbox,
time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, 'BANDS')]
is_data = eopatch[(FeatureType.MASK, 'dataMask')]
self.assertTrue(
np.allclose(array_stats(bands), [0.0233, 0.0468, 0.0252]))
width, height = self.size
self.assertTrue(bands.shape == (4, height, width, 13))
self.assertTrue(is_data.shape == (4, height, width, 1))
self.assertTrue(len(eopatch.timestamp) == 4)
self.assertTrue(bands.dtype == np.float32)
self.assertTrue(os.path.exists(cache_folder))
# change task's bans_dtype and run it again
task.bands_dtype = np.uint16
eopatch = task.execute(bbox=self.bbox,
time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, 'BANDS')]
self.assertTrue(
np.allclose(array_stats(bands), [232.5769, 467.5385, 251.8654]))
self.assertTrue(bands.dtype == np.uint16)
shutil.rmtree(cache_folder)
def test_scl_only(self):
""" Download just SCL, without any other bands
"""
task = SentinelHubInputTask(
bands_feature=None,
additional_data=[(FeatureType.DATA, 'SCL')],
size=self.size,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_collection=DataCollection.SENTINEL2_L2A,
max_threads=self.max_threads
)
eopatch = task.execute(bbox=self.bbox, time_interval=self.time_interval)
scl = eopatch[(FeatureType.DATA, 'SCL')]
width, height = self.size
self.assertTrue(scl.shape == (4, height, width, 1))
def test_aux_request_args(self):
""" Download low resolution data with `PREVIEW` mode
"""
task = SentinelHubInputTask(
bands_feature=(FeatureType.DATA, 'BANDS'),
resolution=260,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_collection=DataCollection.SENTINEL2_L1C,
max_threads=self.max_threads,
aux_request_args={'dataFilter': {'previewMode': 'PREVIEW'}}
)
eopatch = task.execute(bbox=self.bbox, time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, 'BANDS')]
self.assertTrue(bands.shape == (4, 4, 4, 13))
self.assertTrue(np.allclose(array_stats(bands), [0.0, 0.0493, 0.0277]))
def test_aux_request_args(self):
"""Download low resolution data with `PREVIEW` mode"""
task = SentinelHubInputTask(
bands_feature=(FeatureType.DATA, "BANDS"),
resolution=260,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_collection=DataCollection.SENTINEL2_L1C,
max_threads=self.max_threads,
aux_request_args={"dataFilter": {
"previewMode": "PREVIEW"
}},
)
eopatch = task.execute(bbox=self.bbox,
time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, "BANDS")]
assert bands.shape == (4, 4, 4, 13)
assert calculate_stats(bands) == approx([0.0, 0.0493, 0.0277])
def test_specific_bands(self):
task = SentinelHubInputTask(
bands_feature=(FeatureType.DATA, "BANDS"),
bands=["B01", "B02", "B03"],
additional_data=[(FeatureType.MASK, "dataMask")],
size=self.size,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_collection=DataCollection.SENTINEL2_L1C,
max_threads=self.max_threads,
)
eopatch = task.execute(bbox=self.bbox,
time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, "BANDS")]
assert calculate_stats(bands) == approx([0.0648, 0.1193, 0.063])
width, height = self.size
assert bands.shape == (4, height, width, 3)
def test_specific_bands(self):
""" Download S2L1C bands and dataMask
"""
task = SentinelHubInputTask(
bands_feature=(FeatureType.DATA, 'BANDS'),
bands=["B01", "B02", "B03"],
additional_data=[(FeatureType.MASK, 'dataMask')],
size=self.size,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_collection=DataCollection.SENTINEL2_L1C,
max_threads=self.max_threads
)
eopatch = task.execute(bbox=self.bbox, time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, 'BANDS')]
self.assertTrue(np.allclose(array_stats(bands), [0.0648, 0.1193, 0.063]))
width, height = self.size
self.assertTrue(bands.shape == (4, height, width, 3))
def test_single_scene(self):
""" Download S2L1C bands and dataMask
"""
task = SentinelHubInputTask(
bands_feature=(FeatureType.DATA, 'BANDS'),
additional_data=[(FeatureType.MASK, 'dataMask')],
size=self.size,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_collection=DataCollection.SENTINEL2_L1C,
max_threads=self.max_threads,
single_scene=True,
mosaicking_order="leastCC"
)
eopatch = task.execute(bbox=self.bbox, time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, 'BANDS')]
is_data = eopatch[(FeatureType.MASK, 'dataMask')]
width, height = self.size
self.assertTrue(bands.shape == (1, height, width, 13))
self.assertTrue(is_data.shape == (1, height, width, 1))
self.assertTrue(len(eopatch.timestamp) == 1)
def test_s2l1c_float32_uint16(self, cache_folder):
task = SentinelHubInputTask(
bands_feature=(FeatureType.DATA, "BANDS"),
additional_data=[(FeatureType.MASK, "dataMask")],
size=self.size,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_collection=DataCollection.SENTINEL2_L1C,
max_threads=self.max_threads,
cache_folder=cache_folder,
)
eopatch = task.execute(bbox=self.bbox,
time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, "BANDS")]
is_data = eopatch[(FeatureType.MASK, "dataMask")]
assert calculate_stats(bands) == approx([0.0233, 0.0468, 0.0252])
width, height = self.size
assert bands.shape == (4, height, width, 13)
assert is_data.shape == (4, height, width, 1)
assert len(eopatch.timestamp) == 4
assert bands.dtype == np.float32
assert os.path.exists(cache_folder)
# change task's bans_dtype and run it again
task.bands_dtype = np.uint16
eopatch = task.execute(bbox=self.bbox,
time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, "BANDS")]
assert calculate_stats(bands) == approx([232.5769, 467.5385, 251.8654])
assert bands.dtype == np.uint16
def test_S2L2A(self):
""" Download just SCL, without other bands
"""
task = SentinelHubInputTask(bands_feature=(FeatureType.DATA, 'BANDS'),
additional_data=[(FeatureType.MASK,
'dataMask')],
size=self.size,
maxcc=self.maxcc,
time_difference=self.time_difference,
data_source=DataSource.SENTINEL2_L2A,
max_threads=self.max_threads)
eopatch = task.execute(bbox=self.bbox,
time_interval=self.time_interval)
bands = eopatch[(FeatureType.DATA, 'BANDS')]
is_data = eopatch[(FeatureType.MASK, 'dataMask')]
self.assertTrue(
np.allclose(array_stats(bands), [0.0107, 0.0123, 0.0087]))
width, height = self.size
self.assertTrue(bands.shape == (4, height, width, 12))
self.assertTrue(is_data.shape == (4, height, width, 1))
self.assertTrue(len(eopatch.timestamp) == 4)
input_json = json.loads(line)
features = input_json["geometry"]
dam_nominal = shape(features)
dam_bbox = get_bbox(dam_nominal)
# Use login credentials from Sentinel Hub
config = login_config(input_json["access_settings"]["CLIENT_ID"],
input_json["access_settings"]["CLIENT_SECRET"],
input_json["access_settings"]["instance_id"])
# Create an EOPatch and add all EO features (satellite imagery data)
download_task = SentinelHubInputTask(
data_collection=DataCollection.SENTINEL2_L1C,
bands_feature=(FeatureType.DATA, 'BANDS'),
resolution=20,
maxcc=0.5,
bands=['B02', 'B03', 'B04', 'B08'],
additional_data=[(FeatureType.MASK, 'dataMask', 'IS_DATA'),
(FeatureType.MASK, 'CLM')],
config=config)
calculate_ndwi = NormalizedDifferenceIndexTask(
(FeatureType.DATA, 'BANDS'), (FeatureType.DATA, 'NDWI'), (1, 3))
dam_gdf = gpd.GeoDataFrame(crs=CRS.WGS84.pyproj_crs(),
geometry=[dam_nominal])
add_nominal_water = VectorToRaster(
dam_gdf, (FeatureType.MASK_TIMELESS, 'NOMINAL_WATER'),
values=1,
raster_shape=(FeatureType.MASK, 'IS_DATA'),
raster_dtype=np.uint8)
# AOI bounding box
full_bbox = BBox(bbox=[
lon, lat, lon + grid_size * block_width, lat + grid_size * block_height
],
crs=CRS.WGS84)
# ---------------------------------------------------------------------------------------------------------- #
# Cloud-based image filtering to find viable blocks #
# ---------------------------------------------------------------------------------------------------------- #
# eo-learn request
input_task = SentinelHubInputTask(
data_collection=DataCollection.SENTINEL2_L1C,
additional_data=[(FeatureType.DATA, 'CLP'),
(FeatureType.MASK, 'dataMask')],
time_difference=datetime.timedelta(seconds=1),
resolution=60,
config=config)
timelapse = LinearWorkflow(input_task)
try:
result = timelapse.execute(
{input_task: {
'bbox': full_bbox,
'time_interval': time_interval
}})
except:
continue
# parse cloud images and register useful blocks
def setUpClass(cls):
bbox = BBox(bbox=(-5.05, 48.0, -5.00, 48.05), crs=CRS.WGS84)
bbox2 = BBox(bbox=(-72.2, -70.4, -71.6, -70.2), crs=CRS.WGS84)
slo_bbox = BBox([14.5, 45.9, 14.7, 46.1], crs=CRS.WGS84)
cls.size = (50, 40)
time_interval = ('2020-06-1', '2020-06-10')
time_difference = dt.timedelta(minutes=60)
cls.data_feature = FeatureType.DATA, 'BANDS'
cls.mask_feature = FeatureType.MASK, 'dataMask'
s3slstr_500m = DataCollection.SENTINEL3_SLSTR.define_from(
'SENTINEL3_SLSTR_500m',
bands=('S2', 'S3', 'S6')
)
s5p_co = DataCollection.SENTINEL5P.define_from(
'SENTINEL5P_CO',
bands=('CO',)
)
ndvi_evalscript = """
//VERSION=3
function setup() {
return {
input: [{
bands: ["B04", "B08", "dataMask"]
}],
output: [
{ id:"custom", bands:2, sampleType: SampleType.FLOAT32 },
{ id:"bool_mask", bands:1, sampleType: SampleType.UINT8 }
]
}
}
function evaluatePixel(sample) {
return {custom: [index(sample.B08, sample.B04)], bool_mask: [sample.dataMask]};
}
"""
cls.test_cases = [
IoTestCase(
name='Sentinel-2 L2A',
request=SentinelHubInputTask(
bands_feature=cls.data_feature,
additional_data=[cls.mask_feature],
size=cls.size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL2_L2A
),
bbox=bbox,
time_interval=time_interval,
data_size=12,
timestamp_length=2,
stats=[0.4681, 0.6334, 0.7608]
),
IoTestCase(
name='Sentinel-2 L2A - NDVI evalscript',
request=SentinelHubInputTask(
bands_feature=cls.data_feature,
additional_data=[cls.mask_feature],
size=cls.size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL2_L2A,
evalscript=ndvi_evalscript,
bands=['NDVI']
),
bbox=bbox,
time_interval=time_interval,
data_size=1,
timestamp_length=2,
stats=[0.0036, 0.0158, 0.0088]
),
IoTestCase(
name='Landsat8',
request=SentinelHubInputTask(
bands_feature=cls.data_feature,
additional_data=[cls.mask_feature],
size=cls.size,
time_difference=time_difference,
data_collection=DataCollection.LANDSAT8
),
bbox=bbox,
time_interval=time_interval,
data_size=11,
timestamp_length=1,
stats=[0.2206, 0.2654, 0.198]
),
IoTestCase(
name='MODIS',
request=SentinelHubInputTask(
bands_feature=cls.data_feature,
additional_data=[cls.mask_feature],
size=cls.size,
time_difference=time_difference,
data_collection=DataCollection.MODIS
),
bbox=bbox,
time_interval=time_interval,
data_size=7,
timestamp_length=10,
stats=[0.0073, 0.0101, 0.1448]
),
IoTestCase(
name='Sentinel-1 IW',
request=SentinelHubInputTask(
bands_feature=cls.data_feature,
additional_data=[cls.mask_feature],
size=cls.size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL1_IW
),
bbox=bbox,
time_interval=time_interval,
data_size=2,
timestamp_length=5,
stats=[0.0165, 0.0024, 0.0087]
),
IoTestCase(
name='Sentinel-1 IW ASCENDING',
request=SentinelHubInputTask(
bands_feature=cls.data_feature,
additional_data=[cls.mask_feature],
size=cls.size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL1_IW_ASC
),
bbox=bbox,
time_interval=time_interval,
data_size=2,
timestamp_length=1,
stats=[0.0428, 0.0199, 0.022]
),
IoTestCase(
name='Sentinel-1 EW DESCENDING',
request=SentinelHubInputTask(
bands_feature=cls.data_feature,
additional_data=[cls.mask_feature],
size=cls.size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL1_EW_DES
),
bbox=bbox2,
time_interval=time_interval,
data_size=2,
timestamp_length=1,
stats=[np.nan, 0.1919, 0.4114]
),
IoTestCase(
name='Sentinel-3 OLCI',
request=SentinelHubInputTask(
bands_feature=cls.data_feature,
additional_data=[cls.mask_feature],
size=cls.size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL3_OLCI
),
bbox=bbox,
time_interval=time_interval,
data_size=21,
timestamp_length=11,
stats=[0.2064, 0.1354, 0.1905]
),
IoTestCase(
name='Sentinel-3 SLSTR 500m resolution',
request=SentinelHubInputTask(
bands_feature=cls.data_feature,
additional_data=[cls.mask_feature],
size=cls.size,
time_difference=time_difference,
data_collection=s3slstr_500m
),
bbox=bbox,
time_interval=('2021-02-10', '2021-02-15'),
data_size=3,
timestamp_length=14,
stats=[0.4236, 0.6353, 0.5117]
),
IoTestCase(
name='Sentinel-5P',
request=SentinelHubInputTask(
bands_feature=cls.data_feature,
additional_data=[cls.mask_feature],
size=cls.size,
time_difference=time_difference,
data_collection=s5p_co
),
bbox=bbox,
time_interval=('2020-06-1', '2020-06-1'),
data_size=1,
timestamp_length=1,
stats=[0.0351, 0.034, 0.0351]
)
]
class TestSentinelHubInputTaskDataCollections:
"""Integration tests for all supported data collections"""
bbox = BBox(bbox=(-5.05, 48.0, -5.00, 48.05), crs=CRS.WGS84)
bbox2 = BBox(bbox=(-72.2, -70.4, -71.6, -70.2), crs=CRS.WGS84)
size = (50, 40)
time_interval = ("2020-06-1", "2020-06-10")
time_difference = dt.timedelta(minutes=60)
data_feature = FeatureType.DATA, "BANDS"
mask_feature = FeatureType.MASK, "dataMask"
s3slstr_500m = DataCollection.SENTINEL3_SLSTR.define_from(
"SENTINEL3_SLSTR_500m",
bands=(
Band("S2", (Unit.REFLECTANCE, ), (np.float32, )),
Band("S3", (Unit.REFLECTANCE, ), (np.float32, )),
Band("S6", (Unit.REFLECTANCE, ), (np.float32, )),
),
)
s5p_co = DataCollection.SENTINEL5P.define_from("SENTINEL5P_CO",
bands=(Band(
"CO", (Unit.DN, ),
(np.float32, )), ))
ndvi_evalscript = """
//VERSION=3
function setup() {
return {
input: [{
bands: ["B04", "B08", "dataMask"],
units: ["REFLECTANCE", "REFLECTANCE", "DN"]
}],
output: [
{ id:"ndvi", bands:1, sampleType: SampleType.FLOAT32 },
{ id:"dataMask", bands:1, sampleType: SampleType.UINT8 }
]
}
}
function evaluatePixel(sample) {
return {
ndvi: [index(sample.B08, sample.B04)],
dataMask: [sample.dataMask]};
}
"""
test_cases = [
IoTestCase(
name="Sentinel-2 L2A",
task=SentinelHubInputTask(
bands_feature=data_feature,
additional_data=[mask_feature],
size=size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL2_L2A,
),
bbox=bbox,
time_interval=time_interval,
data_size=12,
timestamp_length=2,
stats=[0.4676, 0.6313, 0.7688],
),
IoTestCase(
name="Sentinel-2 L2A - NDVI evalscript",
task=SentinelHubEvalscriptTask(
features={
FeatureType.DATA: [("ndvi", "NDVI-FEATURE")],
FeatureType.MASK: ["dataMask"],
},
evalscript=ndvi_evalscript,
size=size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL2_L2A,
),
feature="NDVI-FEATURE",
bbox=bbox,
time_interval=time_interval,
data_size=1,
timestamp_length=2,
stats=[0.0088, 0.0083, 0.0008],
),
IoTestCase(
name="Landsat8",
task=SentinelHubInputTask(
bands_feature=data_feature,
additional_data=[mask_feature],
size=size,
time_difference=time_difference,
data_collection=DataCollection.LANDSAT_OT_L1,
),
bbox=bbox,
time_interval=time_interval,
data_size=11,
timestamp_length=1,
stats=[48.7592, 48.726, 48.9168],
),
IoTestCase(
name="MODIS",
task=SentinelHubInputTask(
bands_feature=data_feature,
additional_data=[mask_feature],
size=size,
time_difference=time_difference,
data_collection=DataCollection.MODIS,
),
bbox=bbox,
time_interval=time_interval,
data_size=7,
timestamp_length=10,
stats=[0.0073, 0.0101, 0.1448],
),
IoTestCase(
name="Sentinel-1 IW",
task=SentinelHubInputTask(
bands_feature=data_feature,
additional_data=[mask_feature],
size=size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL1_IW,
),
bbox=bbox,
time_interval=time_interval,
data_size=2,
timestamp_length=5,
stats=[0.016, 0.0022, 0.0087],
),
IoTestCase(
name="Sentinel-1 IW ASCENDING",
task=SentinelHubInputTask(
bands_feature=data_feature,
additional_data=[mask_feature],
size=size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL1_IW_ASC,
),
bbox=bbox,
time_interval=time_interval,
data_size=2,
timestamp_length=1,
stats=[0.0407, 0.0206, 0.0216],
),
IoTestCase(
name="Sentinel-1 EW DESCENDING",
task=SentinelHubInputTask(
bands_feature=data_feature,
additional_data=[mask_feature],
size=size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL1_EW_DES,
),
bbox=bbox2,
time_interval=time_interval,
data_size=2,
timestamp_length=1,
stats=[np.nan, 0.1944, 0.3800],
),
IoTestCase(
name="Sentinel-3 OLCI",
task=SentinelHubInputTask(
bands_feature=data_feature,
additional_data=[mask_feature],
size=size,
time_difference=time_difference,
data_collection=DataCollection.SENTINEL3_OLCI,
),
bbox=bbox,
time_interval=time_interval,
data_size=21,
timestamp_length=11,
stats=[0.317, 0.1946, 0.2884],
),
IoTestCase(
name="Sentinel-3 SLSTR 500m resolution",
task=SentinelHubInputTask(
bands_feature=data_feature,
additional_data=[mask_feature],
size=size,
time_difference=time_difference,
data_collection=s3slstr_500m,
),
bbox=bbox,
time_interval=("2021-02-10", "2021-02-15"),
data_size=3,
timestamp_length=13,
stats=[0.3173, 0.4804, 0.4041],
),
IoTestCase(
name="Sentinel-5P",
task=SentinelHubInputTask(
bands_feature=data_feature,
additional_data=[mask_feature],
size=size,
time_difference=time_difference,
data_collection=s5p_co,
),
bbox=bbox,
time_interval=("2020-06-1", "2020-06-1"),
data_size=1,
timestamp_length=1,
stats=[0.0351, 0.034, 0.0351],
),
]
@pytest.mark.parametrize("test_case", test_cases)
def test_data_collections(self, test_case):
eopatch = test_case.task.execute(bbox=test_case.bbox,
time_interval=test_case.time_interval)
assert isinstance(eopatch, EOPatch), "Expected return type is EOPatch"
width, height = self.size
data = eopatch[(test_case.feature_type, test_case.feature)]
assert data.shape == (test_case.timestamp_length, height, width,
test_case.data_size)
timestamps = eopatch.timestamp
assert all(timestamp.tzinfo is None for timestamp in
timestamps), f"`tzinfo` present in timestamps {timestamps}"
assert len(timestamps) == test_case.timestamp_length
stats = calculate_stats(data)
assert stats == approx(
test_case.stats,
nan_ok=True), f"Expected stats {test_case.stats}, got {stats}"
axis=1,
inplace=True)
#SentinelHubInputTask
band_names = []
for band in list(map(int, BANDS.strip('[]').split(','))):
if int(band) < 10:
band_names.append(f'B0{band}')
else:
band_names.append(f'B{band}')
add_data = SentinelHubInputTask(
bands_feature=(FeatureType.DATA, 'BANDS'),
bands=band_names,
resolution=10,
maxcc=MAXCC,
data_collection=DataCollection.SENTINEL2_L1C,
additional_data=[(FeatureType.MASK, 'dataMask'),
(FeatureType.MASK, 'CLM'), (FeatureType.DATA, 'CLP')],
max_threads=5)
#IndexTasks
ndvi = NormalizedDifferenceIndexTask(
(FeatureType.DATA, 'BANDS'), (FeatureType.DATA, 'NDVI'),
[band_names.index('B08'),
band_names.index('B04')])
ndwi = NormalizedDifferenceIndexTask(
(FeatureType.DATA, 'BANDS'), (FeatureType.DATA, 'NDWI'),
[band_names.index('B03'),
band_names.index('B08')])