def _anonymize_files(dicom_directory_in, dicom_directory_out, fields_to_keep):
"""
See anonymize_file for more information.
series_UID and instance_UID will create a new UID respectively for the series for each directory or for the
instance for each file. Note that for a multi-series dataset it is thus required that each series is in its own
directory.
"""
# Make sure we have absolute paths
dicom_directory_in = os.path.abspath(dicom_directory_in)
dicom_directory_out = os.path.abspath(dicom_directory_out)
# looping over all files
for root, _, file_names in os.walk(dicom_directory_in):
# New directory
for file_name in file_names:
# Create instance_UID
fields_to_keep['SOPInstanceUID'] = pydicom.uid.generate_uid()
dicom_file_in = os.path.join(root, file_name)
current_dir = root[len(dicom_directory_in) + 1:]
dicom_file_out = os.path.join(dicom_directory_out, current_dir,
file_name)
if compressed_dicom.is_dicom_file(dicom_file_in):
logging.info("Processing " + dicom_file_in)
_anonymize_file(dicom_file_in, dicom_file_out, fields_to_keep)
else:
logging.info("Skipping " + dicom_file_in + ", no dicom file")
def test_is_dicom_file(self):
input_file = os.path.join(test_data.GENERIC_COMPRESSED, 'IM-0001-0001-0001.dcm')
assert is_dicom_file(input_file)
temporary_directory = tempfile.mkdtemp()
try:
# test for empty file
non_dicom1 = os.path.join(temporary_directory, 'non_dicom.dcm')
open(non_dicom1, 'a').close()
assert not is_dicom_file(non_dicom1)
# test for non empty file
non_dicom2 = os.path.join(temporary_directory, 'non_dicom2.dcm')
with open(non_dicom2, 'w') as file_2:
file_2.write(''.join(random.SystemRandom().choice(string.digits) for _ in range(300)))
assert not is_dicom_file(non_dicom2)
finally:
shutil.rmtree(temporary_directory)
def test_is_dicom_file(self):
input_file = os.path.join(test_data.GENERIC_COMPRESSED,
'IM-0001-0001-0001.dcm')
assert is_dicom_file(input_file)
temporary_directory = tempfile.mkdtemp()
try:
# test for empty file
non_dicom1 = os.path.join(temporary_directory, 'non_dicom.dcm')
open(non_dicom1, 'a').close()
assert not is_dicom_file(non_dicom1)
# test for non empty file
non_dicom2 = os.path.join(temporary_directory, 'non_dicom2.dcm')
with open(non_dicom2, 'w') as file_2:
file_2.write(''.join(
random.SystemRandom().choice(string.digits)
for _ in range(300)))
assert not is_dicom_file(non_dicom2)
finally:
shutil.rmtree(temporary_directory)
def _get_first_header(dicom_directory):
"""
Function to get the first dicom file form a directory and return the header
Useful to determine the type of data to convert
:param dicom_directory: directory with dicom files
"""
# looping over all files
for root, _, file_names in os.walk(dicom_directory):
# go over all the files and try to read the dicom header
for file_name in file_names:
file_path = os.path.join(root, file_name)
# check wither it is a dicom file
if not compressed_dicom.is_dicom_file(file_path):
continue
# read the headers
return compressed_dicom.read_file(file_path,
stop_before_pixels=True,
force=dicom2nifti.settings.pydicom_read_force)
# no dicom files found
raise ConversionError('NO_DICOM_FILES_FOUND')
def read_dicom_directory(dicom_directory, stop_before_pixels=False):
"""
Read all dicom files in a given directory (stop before pixels)
:type stop_before_pixels: bool
:type dicom_directory: six.string_types
:param stop_before_pixels: Should we stop reading before the pixeldata (handy if we only want header info)
:param dicom_directory: Directory with dicom data
:return: List of dicom objects
"""
dicom_input = []
for root, _, files in os.walk(dicom_directory):
for dicom_file in files:
file_path = os.path.join(root, dicom_file)
if compressed_dicom.is_dicom_file(file_path):
dicom_headers = compressed_dicom.read_file(file_path,
defer_size=100,
stop_before_pixels=stop_before_pixels,
force=dicom2nifti.settings.pydicom_read_force)
dicom_input.append(dicom_headers)
return dicom_input
def _get_first_header(dicom_directory):
"""
Function to get the first dicom file form a directory and return the header
Useful to determine the type of data to convert
:param dicom_directory: directory with dicom files
"""
# looping over all files
for root, _, file_names in os.walk(dicom_directory):
# go over all the files and try to read the dicom header
for file_name in file_names:
file_path = os.path.join(root, file_name)
# check wither it is a dicom file
if not compressed_dicom.is_dicom_file(file_path):
continue
# read the headers
return compressed_dicom.read_file(file_path,
stop_before_pixels=True,
force=dicom2nifti.settings.pydicom_read_force)
# no dicom files found
raise ConversionError('NO_DICOM_FILES_FOUND')
def read_dicom_directory(dicom_directory, stop_before_pixels=False):
"""
Read all dicom files in a given directory (stop before pixels)
:type stop_before_pixels: bool
:type dicom_directory: six.string_types
:param stop_before_pixels: Should we stop reading before the pixeldata (handy if we only want header info)
:param dicom_directory: Directory with dicom data
:return: List of dicom objects
"""
dicom_input = []
for root, _, files in os.walk(dicom_directory):
for dicom_file in files:
file_path = os.path.join(root, dicom_file)
if compressed_dicom.is_dicom_file(file_path):
dicom_headers = compressed_dicom.read_file(file_path,
defer_size="1 KB",
stop_before_pixels=stop_before_pixels,
force=dicom2nifti.settings.pydicom_read_force)
if is_valid_imaging_dicom(dicom_headers):
dicom_input.append(dicom_headers)
return dicom_input
def convert_directory(dicom_directory, output_folder, compression=True, reorient=True):
"""
This function will order all dicom files by series and order them one by one
:param compression: enable or disable gzip compression
:param reorient: reorient the dicoms according to LAS orientation
:param output_folder: folder to write the nifti files to
:param dicom_directory: directory with dicom files
"""
# sort dicom files by series uid
dicom_series = {}
for root, _, files in os.walk(dicom_directory):
for dicom_file in files:
file_path = os.path.join(root, dicom_file)
# noinspection PyBroadException
try:
if compressed_dicom.is_dicom_file(file_path):
# read the dicom as fast as possible
# (max length for SeriesInstanceUID is 64 so defer_size 100 should be ok)
dicom_headers = compressed_dicom.read_file(file_path,
defer_size="1 KB",
stop_before_pixels=False,
force=dicom2nifti.settings.pydicom_read_force)
if not _is_valid_imaging_dicom(dicom_headers):
logger.info("Skipping: %s" % file_path)
continue
logger.info("Organizing: %s" % file_path)
if dicom_headers.SeriesInstanceUID not in dicom_series:
dicom_series[dicom_headers.SeriesInstanceUID] = []
dicom_series[dicom_headers.SeriesInstanceUID].append(dicom_headers)
except: # Explicitly capturing all errors here to be able to continue processing all the rest
logger.warning("Unable to read: %s" % file_path)
traceback.print_exc()
# start converting one by one
for series_id, dicom_input in iteritems(dicom_series):
base_filename = ""
# noinspection PyBroadException
try:
# construct the filename for the nifti
base_filename = ""
if 'SeriesNumber' in dicom_input[0]:
base_filename = _remove_accents('%s' % dicom_input[0].SeriesNumber)
if 'SeriesDescription' in dicom_input[0]:
base_filename = _remove_accents('%s_%s' % (base_filename,
dicom_input[0].SeriesDescription))
elif 'SequenceName' in dicom_input[0]:
base_filename = _remove_accents('%s_%s' % (base_filename,
dicom_input[0].SequenceName))
elif 'ProtocolName' in dicom_input[0]:
base_filename = _remove_accents('%s_%s' % (base_filename,
dicom_input[0].ProtocolName))
else:
base_filename = _remove_accents(dicom_input[0].SeriesInstanceUID)
logger.info('--------------------------------------------')
logger.info('Start converting %s' % base_filename)
if compression:
nifti_file = os.path.join(output_folder, base_filename + '.nii.gz')
else:
nifti_file = os.path.join(output_folder, base_filename + '.nii')
convert_dicom.dicom_array_to_nifti(dicom_input, nifti_file, reorient)
gc.collect()
except: # Explicitly capturing app exceptions here to be able to continue processing
logger.info("Unable to convert: %s" % base_filename)
traceback.print_exc()
def convert_directory(dicom_directory, output_folder, compression=True, reorient=True):
"""
This function will order all dicom files by series and order them one by one
:param compression: enable or disable gzip compression
:param reorient: reorient the dicoms according to LAS orientation
:param output_folder: folder to write the nifti files to
:param dicom_directory: directory with dicom files
"""
# sort dicom files by series uid
dicom_series = {}
for root, _, files in os.walk(dicom_directory):
for dicom_file in files:
file_path = os.path.join(root, dicom_file)
# noinspection PyBroadException
try:
if compressed_dicom.is_dicom_file(file_path):
# read the dicom as fast as possible
# (max length for SeriesInstanceUID is 64 so defer_size 100 should be ok)
dicom_headers = compressed_dicom.read_file(file_path,
defer_size="1 KB",
stop_before_pixels=False,
force=dicom2nifti.settings.pydicom_read_force)
if not _is_valid_imaging_dicom(dicom_headers):
logger.info("Skipping: %s" % file_path)
continue
logger.info("Organizing: %s" % file_path)
if dicom_headers.SeriesInstanceUID not in dicom_series:
dicom_series[dicom_headers.SeriesInstanceUID] = []
dicom_series[dicom_headers.SeriesInstanceUID].append(dicom_headers)
except: # Explicitly capturing all errors here to be able to continue processing all the rest
logger.warning("Unable to read: %s" % file_path)
traceback.print_exc()
# start converting one by one
for series_id, dicom_input in dicom_series.items():
base_filename = ""
# noinspection PyBroadException
try:
# construct the filename for the nifti
base_filename = ""
if 'SeriesNumber' in dicom_input[0]:
base_filename = _remove_accents('%s' % dicom_input[0].SeriesNumber)
if 'SeriesDescription' in dicom_input[0]:
base_filename = _remove_accents('%s_%s' % (base_filename,
dicom_input[0].SeriesDescription))
elif 'SequenceName' in dicom_input[0]:
base_filename = _remove_accents('%s_%s' % (base_filename,
dicom_input[0].SequenceName))
elif 'ProtocolName' in dicom_input[0]:
base_filename = _remove_accents('%s_%s' % (base_filename,
dicom_input[0].ProtocolName))
else:
base_filename = _remove_accents(dicom_input[0].SeriesInstanceUID)
logger.info('--------------------------------------------')
logger.info('Start converting %s' % base_filename)
if compression:
nifti_file = os.path.join(output_folder, base_filename + '.nii.gz')
else:
nifti_file = os.path.join(output_folder, base_filename + '.nii')
convert_dicom.dicom_array_to_nifti(dicom_input, nifti_file, reorient)
gc.collect()
except: # Explicitly capturing app exceptions here to be able to continue processing
logger.info("Unable to convert: %s" % base_filename)
traceback.print_exc()
