def store_compressed_volume(self):
if len(self.volume) == 0 or self.volume_compressed:
return
volume = self.volume
compressed_volume_file = os.path.join(
self.compressed_volume_directory,
string2hash(self.uid) + '.dat.gz')
with open(compressed_volume_file, 'wb') as f:
gzip_file = gzip.GzipFile(mode='wb', fileobj=f, compresslevel=9)
gzip_file.write(volume.tobytes())
gzip_file.close()
def fromID(cls,
uid,
scanfolder='',
read_volume=True,
prefer_compressed_volume=True,
compressed_volumes_dir=None):
series_uid = SeriesUID(uid)
scan = cls()
scan_dir = os.path.join(scanfolder, series_uid.dicom_dir_name)
if not os.path.isdir(scan_dir):
scan_dir = os.path.join(scanfolder,
series_uid.optimized_dicom_dir_name)
if not os.path.isdir(scan_dir):
raise IOError("Could not find dicom dir " + scan_dir)
series_builder = SeriesBuilder(uid)
slices = 0
for f in os.listdir(scan_dir):
ds = dicom.read_file(os.path.join(scan_dir, f),
force=True,
stop_before_pixels=not read_volume)
z_pos = get_image_z_position_or_default(ds)
scan.image_positions.append(z_pos)
if read_volume:
series_builder.add_slice(dicom.read_file(
os.path.join(scan_dir, f)),
override_z_pos=z_pos)
slices += 1
scan.size = [slices, ds.Rows, ds.Columns]
scan.uid = uid
scan.directory = scanfolder
scan.compressed_volume_directory = compressed_volumes_dir or scanfolder
scan.name = str(ds.PatientName)
scan.metadata = _read_metadata(ds)
scan.body_part = _get_body_part(ds)
if 'ImageOrientationPatient' in scan.metadata:
scan.plane = get_plane(scan.metadata['ImageOrientationPatient'])
scan.image_orientation = scan.metadata['ImageOrientationPatient']
elif slices == 1:
scan.image_orientation = get_orientation_vector_for_plane_name(
AXIAL)
scan.plane = get_plane(scan.image_orientation)
else:
raise NoImageOrientationError()
scan.date = scan.metadata['StudyDate']
scan.time = scan.metadata['StudyTime']
scan.patient_id = scan.metadata[
'PatientID'] if 'PatientID' in scan.metadata else ''
if read_volume:
scan.volume = None
compressed_volume_file = os.path.join(
scan.compressed_volume_directory,
string2hash(uid) + '.dat.gz')
if prefer_compressed_volume and os.path.isfile(
compressed_volume_file):
try:
with open(compressed_volume_file, 'rb') as f:
scan.volume = f.read()
scan.volume_compressed = True
except Exception:
logger.exception('uid %s: Error reading compressed volume',
uid)
if not scan.volume:
ignore_missing_orientation = False
if slices == 1:
ignore_missing_orientation = True
series = series_builder.build(
ignore_missing_orientation=ignore_missing_orientation)
scan.volume = series.volume
scan.volume_compressed = False
return scan
def fromID(cls,
uid,
scanfolder='',
read_volume=True,
prefer_compressed_volume=True,
compressed_volumes_dir=None):
series_uid = SeriesUID(uid)
if len(scanfolder) == 0:
scanfolder = common.getTMpath('CTscans')
filename = os.path.join(scanfolder, series_uid.mat_filename)
if not os.path.isfile(filename):
return dicom_scan.Scan.fromID(uid, scanfolder, read_volume,
prefer_compressed_volume,
compressed_volumes_dir)
rawScan = h5py.File(filename)
if 'volume' not in rawScan:
raise IOError("Invalid scan in file " + filename)
scan = cls()
scan.uid = uid
scan.directory = scanfolder
scan.compressed_volume_directory = compressed_volumes_dir or scanfolder
scan.name = rawScan['name'].value.tostring().decode('utf-16')
scan.defWindow = rawScan['defWindow'].value
scan.normWindow = rawScan['defWindow'].value + rawScan[
'matlabWindowShift'].value
scan.metadata = _read_metadata(rawScan, rawScan['metadata'])
scan.body_part = _get_scan_type(rawScan)
try:
scan.plane = _get_plane(rawScan)
except Exception as e:
if scan.body_part == 'brain':
# We can assume the scan is axial
logger.warning('uid %s: %s - assuming Axial', uid, str(e))
scan.plane = 'axial'
else:
raise e
scan.image_orientation = scan.metadata[
'ImageOrientationPatient'].tolist()
scan.date = scan.metadata['StudyDate']
scan.time = scan.metadata['StudyTime']
scan.patient_id = scan.metadata[
'PatientID'] if 'PatientID' in scan.metadata else ''
if 'size' not in rawScan:
logger.warning(
'CTscan is missing the "size" field - computing from volume')
# Volume shape is read as (slice, columns, rows). We want to reorder it to (slice, rows, columns)
scan.size = rawScan['volume'].shape
scan.size = [scan.size[0], scan.size[2], scan.size[1]]
else:
# The size is read as (rows, columns, slices). To match the volume, we change it to (slices, rows, colums)
scan.size = rawScan['size'].value.flatten().astype(np.uint16)
scan.size = [scan.size[2], scan.size[0], scan.size[1]]
try:
scan.image_positions = _read_image_positions(
rawScan, scan.size[0], scan.plane)
except Exception as e:
logger.warning('uid %s: %s', uid, str(e))
scan.image_positions = []
if read_volume:
scan.volume = None
compressed_volume_file = os.path.join(
scan.compressed_volume_directory,
string2hash(uid) + '.dat.gz')
if prefer_compressed_volume and os.path.isfile(
compressed_volume_file):
try:
with open(compressed_volume_file, 'rb') as f:
scan.volume = f.read()
scan.volume_compressed = True
except Exception:
logger.exception('uid %s: Error reading compressed volume',
uid)
if not scan.volume:
scan.volume = rawScan['volume']
# Volume shape is read as (slice, columns, rows). We want to reorder it to (slice, rows, columns)
scan.volume = np.transpose(scan.volume, (0, 2, 1))
scan.volume_compressed = False
return scan
def generate_patient_location():
if self.patient_name:
return patientLocations[int(string2hash(self.patient_name)) % len(patientLocations)]
else:
return patientLocations[0]
def load_worklist(self,
user,
worklist_filename,
delete_user_metadata_before_loading=False,
overwrite_findings_metadata=False,
overwrite_volumes=False):
if delete_user_metadata_before_loading:
for work_item in user.worklist:
self.delete_work_item(work_item)
worklist_report = []
try:
uids, worklist_metadata = self.load_worklist_data(
worklist_filename)
except IllegalWorklistFormat as e:
self.logger.error(e)
uids = []
worklist_metadata = None
worklist_report.append('{}'.format(e))
# Collect study information
studies = {}
for uid in uids:
try:
try:
scan = Scan.fromID(uid,
self.ct_scans_dir,
read_volume=False)
except IOError:
self.logger.warning(
'Cannot find CTscan file for uid {}'.format(uid))
worklist_report.append({
uid:
'No scan file {} found'.format(string2hash(uid))
})
continue
if 'StudyInstanceUID' in scan.metadata:
study_uid = scan.metadata['StudyInstanceUID']
else:
self.logger.warning(
'Missing StudyInstanceUID for scan with uid %s - using accession number or '
'series uid instead', uid)
if 'AccessionNumber' in scan.metadata:
study_uid = scan.metadata['AccessionNumber']
else:
study_uid = scan.uid
curr_scan_should_have_findings = self._scan_should_have_findings(
scan, worklist_metadata[study_uid]
if study_uid in worklist_metadata else {})
load_finding_files = user.load_finding_files
if load_finding_files and curr_scan_should_have_findings:
findings_filename = self.get_findings_file(scan.uid)
if not os.path.isfile(findings_filename):
self.logger.warning(
'Cannot locate findings JSON file for scan with uid "%s" - skipping',
uid)
worklist_report.append({uid: 'No findings file'})
continue
if study_uid not in studies:
studies[study_uid] = []
# check for duplicate series uids in the worklist
duplicate_found = False
for added_scan in studies[study_uid]:
if added_scan.uid == scan.uid:
duplicate_found = True
break
if duplicate_found:
self.logger.warning(
'Duplicate uid {} in worklist'.format(uid))
worklist_report.append({uid: 'Duplicate uid in worklist'})
continue
studies[study_uid].append(scan)
except Exception as e:
self.logger.warning(
'Unknown error for uid {} string2hash {}'.format(
uid, string2hash(uid)))
worklist_report.append({
uid:
'Unknown error for uid {} string2hash {}'.format(
uid, string2hash(uid))
})
study_count = 0
study_load_count = 0
series_count = 0
series_load_count = 0
studies_report = []
for study_uid, scans in studies.items():
try:
loaded, loaded_series = self.load_study(
study_uid, scans, user.id, worklist_metadata[study_uid] if
study_uid in worklist_metadata else {}, overwrite_volumes,
load_finding_files, overwrite_findings_metadata)
study_count += 1
series_count += len(scans)
if loaded:
study_load_count += 1
series_load_count += loaded_series
except Exception as e:
self.logger.warning(str(e))
studies_report.append({'study': study_uid, 'error': str(e)})
return {
'worklist_report': {
'loaded': series_load_count,
'already-exist': series_count - series_load_count,
'errors': worklist_report
},
'studies_report': {
'loaded': study_load_count,
'already-exist': study_count - study_load_count,
'errors': studies_report
}
}
from dicomaizer.common.uid import string2hash worklist_file = '/Users/annabankirer/Downloads/sheba_abdomen' command = ' scp -i ~/.ssh/aidoc [email protected]:/datadrive/viewer-data/scans/{}.mat .' with open(worklist_file) as f: for line in f: line = line.strip() if len(line) == 0: continue print(command.format(string2hash(line)))