def test_dec10216(self):
res = dec10216(np.array([255, 255, 255, 255, 255], dtype=np.uint8))
exp = (np.ones((4, )) * 1023).astype(np.uint16)
self.assertTrue(np.all(res == exp))
res = dec10216(np.array([1, 1, 1, 1, 1], dtype=np.uint8))
exp = np.array([4, 16, 64, 257], dtype=np.uint16)
self.assertTrue(np.all(res == exp))
def get_dataset(self, dsid, info,
xslice=slice(None), yslice=slice(None)):
if dsid.name not in self.mda['channel_list']:
raise KeyError('Channel % s not available in the file' % dsid.name)
elif dsid.name not in ['HRV']:
shape = (self.mda['number_of_lines'], self.mda['number_of_columns'])
# Check if there is only 1 channel in the list as a change
# is needed in the arrray assignment ie channl id is not present
if len(self.mda['channel_list']) == 1:
raw = self.dask_array['visir']['line_data']
else:
i = self.mda['channel_list'].index(dsid.name)
raw = self.dask_array['visir']['line_data'][:, i, :]
data = dec10216(raw.flatten())
data = da.flipud(da.fliplr((data.reshape(shape))))
else:
shape = (self.mda['hrv_number_of_lines'], self.mda['hrv_number_of_columns'])
raw2 = self.dask_array['hrv']['line_data'][:, 2, :]
raw1 = self.dask_array['hrv']['line_data'][:, 1, :]
raw0 = self.dask_array['hrv']['line_data'][:, 0, :]
shape_layer = (self.mda['number_of_lines'], self.mda['hrv_number_of_columns'])
data2 = dec10216(raw2.flatten())
data2 = da.flipud(da.fliplr((data2.reshape(shape_layer))))
data1 = dec10216(raw1.flatten())
data1 = da.flipud(da.fliplr((data1.reshape(shape_layer))))
data0 = dec10216(raw0.flatten())
data0 = da.flipud(da.fliplr((data0.reshape(shape_layer))))
data = np.zeros(shape)
idx = range(0, shape[0], 3)
data[idx, :] = data2
idx = range(1, shape[0], 3)
data[idx, :] = data1
idx = range(2, shape[0], 3)
data[idx, :] = data0
xarr = xr.DataArray(data, dims=['y', 'x']).where(data != 0).astype(np.float32)
if xarr is None:
dataset = None
else:
dataset = self.calibrate(xarr, dsid)
dataset.attrs['units'] = info['units']
dataset.attrs['wavelength'] = info['wavelength']
dataset.attrs['standard_name'] = info['standard_name']
dataset.attrs['platform_name'] = self.mda['platform_name']
dataset.attrs['sensor'] = 'seviri'
dataset.attrs['orbital_parameters'] = {
'projection_longitude': self.mda['projection_parameters']['ssp_longitude'],
'projection_latitude': 0.,
'projection_altitude': self.mda['projection_parameters']['h']}
return dataset
def test_dec10216(self):
"""Test the dec10216 function."""
res = dec10216(np.array([255, 255, 255, 255, 255], dtype=np.uint8))
exp = (np.ones((4, )) * 1023).astype(np.uint16)
np.testing.assert_equal(res, exp)
res = dec10216(np.array([1, 1, 1, 1, 1], dtype=np.uint8))
exp = np.array([4, 16, 64, 257], dtype=np.uint16)
np.testing.assert_equal(res, exp)
def _get_visir_channel(self, dataset_id):
shape = (self.mda['number_of_lines'], self.mda['number_of_columns'])
# Check if there is only 1 channel in the list as a change
# is needed in the arrray assignment ie channl id is not present
if len(self.mda['channel_list']) == 1:
raw = self.dask_array['visir']['line_data']
else:
i = self.mda['channel_list'].index(dataset_id['name'])
raw = self.dask_array['visir']['line_data'][:, i, :]
data = dec10216(raw.flatten())
data = data.reshape(shape)
return data
def _get_hrv_channel(self):
shape = (self.mda['hrv_number_of_lines'],
self.mda['hrv_number_of_columns'])
shape_layer = (self.mda['number_of_lines'],
self.mda['hrv_number_of_columns'])
data_list = []
for i in range(3):
raw = self.dask_array['hrv']['line_data'][:, i, :]
data = dec10216(raw.flatten())
data = data.reshape(shape_layer)
data_list.append(data)
return np.stack(data_list, axis=1).reshape(shape)
def read_band(self, key, info):
"""Read the data."""
shape = int(np.ceil(self.mda['data_field_length'] / 8.))
if self.mda['number_of_bits_per_pixel'] == 16:
dtype = '>u2'
shape //= 2
elif self.mda['number_of_bits_per_pixel'] in [8, 10]:
dtype = np.uint8
shape = (shape, )
data = self._read_or_memmap_data(shape, dtype)
data = da.from_array(data, chunks=shape[0])
if self.mda['number_of_bits_per_pixel'] == 10:
data = dec10216(data)
data = data.reshape((self.mda['number_of_lines'],
self.mda['number_of_columns']))
return data
def read_band(self, key, info):
"""Read the data."""
shape = int(np.ceil(self.mda['data_field_length'] / 8.))
if self.mda['number_of_bits_per_pixel'] == 16:
dtype = '>u2'
shape //= 2
elif self.mda['number_of_bits_per_pixel'] in [8, 10]:
dtype = np.uint8
shape = (shape, )
# For reading the image data, unzip_context is faster than generic_open
with utils.unzip_context(self.filename) as fn:
data = np.memmap(fn,
mode='r',
offset=self.mda['total_header_length'],
dtype=dtype,
shape=shape)
data = da.from_array(data, chunks=shape[0])
if self.mda['number_of_bits_per_pixel'] == 10:
data = dec10216(data)
data = data.reshape(
(self.mda['number_of_lines'], self.mda['number_of_columns']))
return data
def get_dataset(self, dataset_id, dataset_info):
"""Get the dataset."""
if dataset_id['name'] not in self.mda['channel_list']:
raise KeyError('Channel % s not available in the file' %
dataset_id['name'])
elif dataset_id['name'] not in ['HRV']:
shape = (self.mda['number_of_lines'],
self.mda['number_of_columns'])
# Check if there is only 1 channel in the list as a change
# is needed in the arrray assignment ie channl id is not present
if len(self.mda['channel_list']) == 1:
raw = self.dask_array['visir']['line_data']
else:
i = self.mda['channel_list'].index(dataset_id['name'])
raw = self.dask_array['visir']['line_data'][:, i, :]
data = dec10216(raw.flatten())
data = data.reshape(shape)
else:
shape = (self.mda['hrv_number_of_lines'],
self.mda['hrv_number_of_columns'])
raw2 = self.dask_array['hrv']['line_data'][:, 2, :]
raw1 = self.dask_array['hrv']['line_data'][:, 1, :]
raw0 = self.dask_array['hrv']['line_data'][:, 0, :]
shape_layer = (self.mda['number_of_lines'],
self.mda['hrv_number_of_columns'])
data2 = dec10216(raw2.flatten())
data2 = data2.reshape(shape_layer)
data1 = dec10216(raw1.flatten())
data1 = data1.reshape(shape_layer)
data0 = dec10216(raw0.flatten())
data0 = data0.reshape(shape_layer)
data = np.stack((data0, data1, data2), axis=1).reshape(shape)
xarr = xr.DataArray(data,
dims=['y',
'x']).where(data != 0).astype(np.float32)
if xarr is None:
dataset = None
else:
dataset = self.calibrate(xarr, dataset_id)
dataset.attrs['units'] = dataset_info['units']
dataset.attrs['wavelength'] = dataset_info['wavelength']
dataset.attrs['standard_name'] = dataset_info['standard_name']
dataset.attrs['platform_name'] = self.mda['platform_name']
dataset.attrs['sensor'] = 'seviri'
dataset.attrs['orbital_parameters'] = {
'projection_longitude':
self.mda['projection_parameters']['ssp_longitude'],
'projection_latitude':
0.,
'projection_altitude':
self.mda['projection_parameters']['h']
}
return dataset
