def load_motl(path_to_dataset: str) -> np.array:
_, data_file_extension = os.path.splitext(path_to_dataset)
assert data_file_extension in [".em", ".csv"], "file in non valid format."
if data_file_extension == ".em":
em_header, motl = read_em(path_to_emfile=path_to_dataset)
else: # data_file_extension == ".csv":
motl = read_motl_from_csv(path_to_csv_motl=path_to_dataset)
return motl
def load_tomogram(path_to_dataset: str, dtype=None) -> np.array:
"""
Verified that they open according to same coordinate system
"""
_, data_file_extension = os.path.splitext(path_to_dataset)
print("file in {} format".format(data_file_extension))
assert data_file_extension in [".em", ".hdf", ".mrc", ".rec"], \
"file in non valid format."
if data_file_extension == ".em":
em_header, dataset = read_em(path_to_emfile=path_to_dataset)
elif data_file_extension == ".hdf":
dataset = _load_hdf_dataset(hdf_file_path=path_to_dataset)
elif data_file_extension in [".mrc", ".rec"]:
dataset = read_mrc(path_to_mrc=path_to_dataset, dtype=dtype)
return dataset
def load_motl_as_df(path_to_motl):
_, data_file_extension = os.path.splitext(path_to_motl)
column_names = [
'score', 'x_', 'y_', 'peak', 'tilt_x', 'tilt_y', 'tilt_z', 'x', 'y',
'z', 'empty_1', 'empty_2', 'empty_3', 'x-shift', 'y-shift', 'z-shift',
'phi', 'psi', 'theta', 'class'
]
assert data_file_extension in [".em", ".csv"], "file in non valid format."
if data_file_extension == ".em":
header, motl = read_em(path_to_emfile=path_to_motl)
motl_df = pd.DataFrame(motl, columns=column_names)
else:
motl_df = pd.read_csv(path_to_motl, header=None)
motl_df.columns = column_names
return motl_df
def read_motl_coordinates_and_values(path_to_motl: str) -> tuple:
_, motl_extension = os.path.splitext(path_to_motl)
assert motl_extension in [
".em", ".csv"
], "motl clean should be in a valid format .em or .csv"
if motl_extension == ".em":
print("motl in .em format")
header, motl = read_em(path_to_emfile=path_to_motl)
motl_values, motl_coords = extract_coordinates_and_values_from_em_motl(
motl)
else:
print("motl in .csv format")
motl = read_motl_from_csv(path_to_motl)
motl_values, motl_coords = extract_motl_coordinates_and_score_values(
motl)
motl_coords = np.array(motl_coords, dtype=int)
return motl_values, motl_coords
def generate_particle_mask_from_motl(path_to_motl: str,
output_shape: tuple,
sphere_radius=8,
values_in_motl: bool = True,
number_of_particles=None,
z_shift=0,
particles_in_tom_format=True) -> None:
_, motl_extension = os.path.splitext(path_to_motl)
assert motl_extension in [".csv", ".em"]
if motl_extension == ".csv" or motl_extension == ".em":
if motl_extension == ".csv":
motive_list = read_motl_from_csv(path_to_motl)
if isinstance(number_of_particles, int):
motive_list = motive_list[:number_of_particles]
print("Only", str(number_of_particles),
" particles in the motive list will be pasted.")
else:
print("All particles in the motive list will be pasted.")
if particles_in_tom_format:
coordinates = [
np.array([int(row[9]) + z_shift,
int(row[8]),
int(row[7])]) for row in motive_list
]
else:
coordinates = [
np.array([int(row[7]) + z_shift,
int(row[8]),
int(row[9])]) for row in motive_list
]
if values_in_motl:
score_values = [row[0] for row in motive_list]
else:
score_values = np.ones(len(motive_list))
print("The map will be binary.")
else:
_, motive_list = read_em(path_to_emfile=path_to_motl)
if isinstance(number_of_particles, int):
motive_list = motive_list[:number_of_particles]
print("Only", str(number_of_particles),
" particles in the motive list will be pasted.")
else:
print("All particles in the motive list will be pasted.")
coordinates = extract_coordinates_from_em_motl(motive_list)
if particles_in_tom_format:
print("coordinates already in tom format")
coordinates = [[int(p[2]) + z_shift,
int(p[1]),
int(p[0])] for p in coordinates]
else:
print("transforming coordinates to tom format")
coordinates = [[int(p[2]) + z_shift,
int(p[1]),
int(p[0])] for p in coordinates]
score_values = np.ones(len(coordinates))
predicted_dataset = np.zeros(output_shape)
for center, value in zip(coordinates, score_values):
paste_sphere_in_dataset(dataset=predicted_dataset,
center=center,
radius=sphere_radius,
value=value)
return predicted_dataset