def build_offsets(self):
if self.get("offsets_file_name") is not None:
print("Loading offsets from file...")
offset_dir = self.get(
"offsets_dir_path",
os.path.join(
get_abailoni_hci_home_path(),
"pyCharm_projects/uppsala_hackathon/experiments/cremi/offsets"
))
offsets_path = os.path.join(offset_dir,
self.get("offsets_file_name"))
assert os.path.exists(offsets_path)
with open(offsets_path, 'r') as f:
self.set("offsets", json.load(f))
else:
print("Loading offsets from inference config...")
offset_path_in_infer_config = self.get(
"offset_path_in_infer_config", ensure_exists=True)
affs_dir_path = self.get("affinities_dir_path", ensure_exists=True)
prediction_config = segm_utils.yaml2dict(
os.path.join(affs_dir_path, "prediction_config.yml"))
# Recursively look into the prediction_config:
paths = offset_path_in_infer_config.split("/")
data = prediction_config
for path in paths:
assert path in data
data = data.get(path)
offsets = data
assert isinstance(offsets, list)
self.set("offsets", offsets)
def __init__(self, experiment_directory=None, config=None):
super(PostProcessingExperiment, self).__init__(experiment_directory)
# Privates
self._meta_config['exclude_attrs_from_save'] = [
'data_loader', '_device'
]
if config is not None:
self.read_config_file(config)
self.auto_setup()
# Where the segmentation will be saved: (each experiment is in a sub-folder)
proj_dir_path = self.get(
'proj_dir_path',
os.path.join(get_trendytukan_drive_path(),
"projects/pixel_embeddings"))
self.set("proj_dir_path", proj_dir_path)
# Load the file with all the presets of postprocessing:
postproc_presets_path = self.get(
'postproc_presets_path',
os.path.join(
get_abailoni_hci_home_path(),
"pyCharm_projects/uppsala_hackathon/experiments/cremi/"
"postproc_configs/postproc_presets.yml"))
self.set("postproc_presets",
segm_utils.yaml2dict(postproc_presets_path))
self.build_experiment()
self.build_offsets()
def load_offsets():
offsets_path = os.path.join(
get_abailoni_hci_home_path(),
"pyCharm_projects/uppsala_hackathon/experiments/cremi/offsets",
offsets_file_name)
assert os.path.exists(offsets_path)
with open(offsets_path, 'r') as f:
return json.load(f)
self.dump_configuration(os.path.join(dir_path,
"prediction_config.yml"))
if __name__ == '__main__':
print(sys.argv[1])
source_path = os.path.dirname(os.path.realpath(__file__))
config_path = os.path.join(source_path, 'configs')
experiments_path = os.path.join(source_path, 'runs')
# Update HCI_HOME paths:
for i, key in enumerate(sys.argv):
if "HCI__HOME" in sys.argv[i]:
sys.argv[i] = sys.argv[i].replace("HCI__HOME/",
get_abailoni_hci_home_path())
# Update RUNS paths:
for i, key in enumerate(sys.argv):
if "RUNS__HOME" in sys.argv[i]:
sys.argv[i] = sys.argv[i].replace("RUNS__HOME", experiments_path)
sys.argv[1] = os.path.join(experiments_path, sys.argv[1])
if '--inherit' in sys.argv:
i = sys.argv.index('--inherit') + 1
if sys.argv[i].endswith(('.yml', '.yaml')):
sys.argv[i] = change_paths_config_file(
os.path.join(config_path, sys.argv[i]))
else:
sys.argv[i] = os.path.join(experiments_path, sys.argv[i])
if '--update' in sys.argv:
from matplotlib import pyplot as plt
sample = "C"
# slices = {'A': ":,:,:,:", 'B': ":, :, 90:, 580: 1900", 'C': ":, :, 70:1450, 95:1425"}
slices = {
'A': ":,:,:,:",
'B': ":,:,:,:",
'C': ":,70:-6,50:-50,50:-50",
"0": ":,:,:,:",
"1": ":,:,:,:",
"2": ":,:,:,:",
"3": ":,:,:,:",
}
parsed_slice = parse_data_slice(slices[sample])
data_path = os.path.join(get_abailoni_hci_home_path(),
"datasets/new_cremi/sample{}.h5".format(sample))
# data_path = os.path.join(get_trendytukan_drive_path(), "datasets/new_cremi/fib25/sample{}.h5".format(sample))
with h5py.File(data_path, 'r') as f:
print([atr for atr in f['volumes/labels']])
# glia = f['volumes/labels/glia'][:]
raw = f['volumes/raw'][parsed_slice[1:]]
GT = f['volumes/labels/neuron_ids_fixed'][parsed_slice[1:]]
# Load affs:
from segmfriends.utils.various import writeHDF5, readHDF5
affs_path = os.path.join(
get_trendytukan_drive_path(),
"projects/pixel_embeddings/{}/predictions_sample_{}.h5".format(
"v4_addSparseAffs_eff", sample))
affs_dice = 1. - readHDF5(
active_nodes = np.empty(self.shape, dtype='int64')
for n in range(self.nb_nodes):
out_segm[tuple(self.from_label_to_coord(
n))] = self.c_graph.findRepresentativeNode(n)
active_nodes[tuple(
self.from_label_to_coord(n))] = self.is_node_active[n]
active_nodes[tuple(self.starting_coordinate)] = 2
return out_segm, active_nodes
import os
from vaeAffs.utils.path_utils import get_abailoni_hci_home_path
import h5py
sample = "A"
data_path = os.path.join(
get_abailoni_hci_home_path(),
"../ialgpu1_local_home/datasets/cremi/SOA_affinities/sample{}_train.h5".
format(sample))
crop_slice = ":,20:21,300:400,300:400"
from segmfriends.utils.various import parse_data_slice
crop_slice = parse_data_slice(crop_slice)
with h5py.File(data_path, 'r') as f:
affs = f['predictions']['full_affs'][crop_slice]
raw = f['raw'][crop_slice[1:]]
offsets = [[-1, 0, 0], [0, -1, 0], [0, 0, -1], [-2, 0, 0], [0, -3, 0],
[0, 0, -3], [-3, 0, 0], [0, -9, 0], [0, 0, -9], [-4, 0, 0],
[0, -27, 0], [0, 0, -27]]
# Fake duplicate affinities:
import vaeAffs
import nifty
import numpy as np
import vigra
import os
from vaeAffs.utils.path_utils import get_abailoni_hci_home_path
import h5py
sample = "A"
data_path = os.path.join(
get_abailoni_hci_home_path(),
"../ialgpu1_local_home/datasets/cremi/SOA_affinities/sample{}_train.h5".
format(sample))
crop_slice = ":,20:30,300:600,300:600"
from segmfriends.utils.various import parse_data_slice
crop_slice = parse_data_slice(crop_slice)
with h5py.File(data_path, 'r') as f:
affs = f['predictions']['full_affs'][crop_slice]
raw = f['raw'][crop_slice[1:]]
offsets = [[-1, 0, 0], [0, -1, 0], [0, 0, -1], [-2, 0, 0], [0, -3, 0],
[0, 0, -3], [-3, 0, 0], [0, -9, 0], [0, 0, -9], [-4, 0, 0],
[0, -27, 0], [0, 0, -27]]
from affogato.affinities import compute_multiscale_affinities, compute_affinities
from affogato.segmentation import compute_mws_segmentation
_, mask = compute_affinities(np.zeros_like(raw, dtype='int64'), offsets)
