class TestDownload(unittest.TestCase):
def setUp(self):
self.oo = OCP()
def test_get_cutout_type(self):
# kasthuri11/image/xy/3/1000,1100/1000,1100/1000/
numpy_download = self.oo.get_cutout('kasthuri11', 'image',
1000, 1100,
1000, 1100,
1000, 1001,
resolution=3)
self.assertEqual(type(numpy_download), numpy.ndarray)
def test_get_cutout(self):
# kasthuri11/image/xy/3/1000,1100/1000,1100/1000/
numpy_download = self.oo.get_cutout('kasthuri11', 'image',
1000, 1100,
1000, 1100,
1000, 1001,
resolution=3)
# We know this pixel is 132
self.assertEqual(numpy_download[0][0][0], 132)
def test_get_volume_type(self):
# kasthuri11/image/xy/3/1000,1100/1000,1100/1000/
ramon_download = self.oo.get_volume('kasthuri11', 'image',
1000, 1100,
1000, 1100,
1000, 1001,
resolution=3)
self.assertEqual(type(ramon_download), ndio.ramon.RAMONVolume)
def test_get_volume(self):
# kasthuri11/image/xy/3/1000,1100/1000,1100/1000/
ramon_download = self.oo.get_volume('kasthuri11', 'image',
1000, 1100,
1000, 1100,
1000, 1001,
resolution=3)
self.assertEqual(ramon_download.xyz_offset[0], 1000)
self.assertEqual(ramon_download.resolution, 3)
self.assertEqual(ramon_download.cutout[0][0][0], 132)
def test_get_image_type(self):
# kasthuri11/image/xy/3/1000,1100/1000,1100/1000/
image_download = self.oo.get_image('kasthuri11', 'image',
1000, 1100,
1000, 1100,
1000,
resolution=3)
self.assertEqual(type(image_download), numpy.ndarray)
self.assertEqual(image_download.shape, (100, 100))
def getTruth(db = False):
# Careful of hardcoded filepath
if db == False:
vol = np.array([])
data = ndio.convert.png.import_png_collection('newdata/mito_anno_*')
for img in data:
vol = np.dstack([vol, img]) if vol.size else img
return vol
else:
boundaries = getConfig()
oo = OCP()
#added +12 because first 10 images in stack aren't even annotated
return oo.get_cutout('kasthuri2015_ramon_v1', 'mitochondria', 694 + boundaries[0], 694 + boundaries[1], 1750 + boundaries[2], 1750 + boundaries[3], 1004, 1154, resolution = 3)
def getRaw(x1,x2,y1,y2,z1,z2):
"""
Retrieve mitochondria image volume from neurodata given x,y,z parameters
Arguments:
x1/y1/z1: starting value for cut
x2/y2/z2: ending cut value
Returns:
3D Numpy Array of images (each image is in a 2D numpy array)
"""
oo = OCP()
return oo.get_cutout('kasthuri11cc', 'image', x1, x2, y1, y2, z1, z2, resolution = 3)
# Get all tokens
tokens = oo.get_public_tokens()
# Membrane group image and annotation tokens for membrane image in datamap
image_token = 'kasthuri11cc'
annotation_token = 'cv_kasthuri11_membrane_2014'
# Get channel ROI for token
channel_ROI = nn.get_metadata('cv_kasthuri11_membrane_2014')['channels']
# Get membrane ROI coordinates
membrane_group_ROI = channel_ROI['image']['rois']['ac4']
# Sets membrane query in Python
membrane_query = {
'token': 'cv_kasthuri11_membrane_2014',
'channel': 'image',
'x_start': membrane_group_ROI['x'][0],
'x_stop': membrane_group_ROI['x'][1],
'y_start': membrane_group_ROI['y'][0],
'y_stop': membrane_group_ROI['y'][1],
'z_start': membrane_group_ROI['z'][0],
'z_stop': membrane_group_ROI['z'][1],
'resolution': membrane_group_ROI['resolution'],
}
# Sets cutout arguments
membrane_images = oo.get_cutout(**membrane_query)
ndtiff.export_tiff('membrane_pic1.tiff', membrane_images[0])
annotation_token = "kasthuri2015_ramon_v"
segmentation_token = "ac3ac4"
# Get channel ROI for token
channel_ROI = nn.get_metadata(segmentation_token)["channels"]
# Get membrane ROI coordinates
membrane_group_ROI = channel_ROI["ac4_neuron_truth"]["rois"]["ac4"]
# Sets membrane query in Python
membrane_query = {
"token": "ac3ac4",
"channel": "ac4_neuron_truth",
"x_start": membrane_group_ROI["x"][0],
"x_stop": membrane_group_ROI["x"][1],
"y_start": membrane_group_ROI["y"][0],
"y_stop": membrane_group_ROI["y"][1],
"z_start": membrane_group_ROI["z"][0],
"z_stop": membrane_group_ROI["z"][1],
"resolution": membrane_group_ROI["resolution"],
}
print "Getting images"
# Sets cutout arguments for ground truth membrane annotation
membrane_anno = oo.get_cutout(**membrane_query)
print "Printing to file"
np.save("nXp_ground.npy", membrane_anno)
print "Done."
annotation_token = 'kasthuri2015_ramon_v'
segmentation_token = 'ac3ac4'
# Get channel ROI for token
channel_ROI = nn.get_metadata(segmentation_token)['channels']
# Get membrane ROI coordinates
membrane_group_ROI = channel_ROI['ac4_neuron_truth']['rois']['ac4']
# Sets membrane query in Python
membrane_query = {
'token': 'ac3ac4',
'channel': 'ac4_neuron_truth',
'x_start': membrane_group_ROI['x'][0],
'x_stop': membrane_group_ROI['x'][1],
'y_start': membrane_group_ROI['y'][0],
'y_stop': membrane_group_ROI['y'][1],
'z_start': membrane_group_ROI['z'][0],
'z_stop': membrane_group_ROI['z'][1],
'resolution': membrane_group_ROI['resolution'],
}
print "Getting images"
# Sets cutout arguments for ground truth membrane annotation
membrane_anno = oo.get_cutout(**membrane_query)
print "Printing to file"
np.save('nXp_ground.npy', membrane_anno)
print "Done."
