def _update_deidentification_method(self, dataset: pydicom.dataset.Dataset) -> None:
if "DeidentificationMethod" not in dataset:
dataset.DeidentificationMethod = "DICOGNITO"
return
existing_element = dataset.data_element("DeidentificationMethod")
assert existing_element is not None # satisfy mypy
existing_value = existing_element.value
if isinstance(existing_value, pydicom.multival.MultiValue):
if "DICOGNITO" not in existing_value:
existing_value.append("DICOGNITO")
elif existing_value != "DICOGNITO":
existing_element.value = [existing_value, "DICOGNITO"]
def anonymize(self, dataset: pydicom.dataset.Dataset) -> None:
"""\
Anonymize a dataset in place. Replaces all PNs, UIs, dates and times, and
known identifiying attributes with other vlaues.
Parameters
----------
dataset : pydicom.dataset.Dataset
A DICOM dataset to anonymize.
"""
dataset.file_meta.walk(self._anonymize_element)
dataset.walk(self._anonymize_element)
self._update_deidentification_method(dataset)
self._update_patient_identity_removed(dataset)
def _set_study_attributes(dataset: pydicom.dataset.Dataset, patient_number: int, study_number: int) -> None:
dataset.StudyID = "STUDYFOR4MR" + str(patient_number) + "." + str(study_number)
dataset.AccessionNumber = "ACC" + dataset.StudyID
dataset.StudyDate = datetime.date(2004, patient_number, study_number).strftime("%Y%m%d")
dataset.StudyTime = datetime.time(patient_number * 5 + study_number, 0, 0).strftime("%H%M%S")
dataset.StudyInstanceUID = "1.3.6.1.4.1.5962.20040827145012.5458." + str(patient_number) + "." + str(study_number)
dataset.NameOfPhysiciansReadingStudy = "READING^FIRST^" + dataset.StudyID
dataset.RequestingPhysician = "REQUESTING1^FIRST^" + dataset.StudyID
dataset.ReferringPhysicianName = "REFERRING1^FIRST^" + dataset.StudyID
def anonymize_institution_name(self, dataset: pydicom.dataset.Dataset,
data_element: pydicom.DataElement) -> None:
region = self.address_anonymizer.get_region(data_element.value)
street_address = self.address_anonymizer.get_street_address(
data_element.value)
street = street_address.split(" ", 1)[1]
dataset.InstitutionAddress = ", ".join([
street_address, region,
self.address_anonymizer.get_country(data_element.value)
])
data_element.value = region + "'S " + street + " CLINIC"
def _anonymize_date_and_time(self, dataset: pydicom.dataset.Dataset, data_element: pydicom.DataElement) -> None:
date_value = data_element.value
if isinstance(data_element.value, pydicom.multival.MultiValue):
dates = list([v for v in data_element.value])
else:
dates = [data_element.value]
times = []
time_name = data_element.keyword.replace("Date", "Time")
if time_name in dataset:
time_element = dataset.data_element(time_name)
time_value = time_element.value # type: ignore[union-attr]
if time_value:
if isinstance(time_value, pydicom.multival.MultiValue):
times = list([v for v in time_value])
else:
times = [time_value]
new_dates = []
new_times = []
for i in range(len(dates)):
date_value = dates[i]
date_format = "%Y%m%d"[: len(date_value) - 2]
old_date = datetime.datetime.strptime(date_value, date_format).date()
time_value = ""
old_hours = datetime.time()
if i < len(times):
time_value = times[i]
if time_value:
old_hours = datetime.datetime.strptime(time_value[:2], "%H").time()
else:
old_hours = datetime.time()
old_datetime = datetime.datetime.combine(old_date, old_hours)
new_datetime = old_datetime + self.offset
new_dates.append(new_datetime.strftime(date_format))
new_times.append(new_datetime.strftime("%H") + time_value[2:])
new_dates_string = "\\".join(new_dates)
new_times_string = "\\".join(new_times)
data_element.value = new_dates_string
if times:
time_element.value = new_times_string # type: ignore[union-attr]
def get_new_ds(ds: pydicom.dataset.Dataset, name: str):
"""
Create a new dicom header. Creates a new dataset object and sets all existing
fields in ds.
:param ds: old dataset to base this new one off of
:param name: name for this file
:return: The new dicom dataset object with all header information. No pixel data.
"""
file_meta = ds.file_meta
new_ds = pydicom.dataset.FileDataset(
name, {},
file_meta=file_meta,
preamble=b"\0" * 128,
is_implicit_VR=ds.is_implicit_VR,
is_little_endian=ds.is_little_endian)
a = [
attr for attr in dir(ds)
if not attr.startswith('__') and not callable(getattr(ds, attr))
]
for attr in a:
if attr in [
'_character_set', 'is_original_encoding', 'pixel_array',
'BitsAllocated', 'BitsStored', 'HighBit',
'PixelRepresentation', 'PixelData'
]:
continue
else:
new_ds.__setattr__(attr, ds.__getattr__(attr))
# fix bits
new_ds.BitsAllocated = 8
new_ds.BitsStored = 8
new_ds.HighBit = 7
new_ds.PixelRepresentation = 0
return new_ds
def _set_series_attributes(
dataset: pydicom.dataset.Dataset, patient_number: int, study_number: int, series_number: int
) -> None:
series_suffix = "%(patient_number)-d%(study_number)-d%(series_number)d" % vars()
dataset.SeriesInstanceUID = dataset.StudyInstanceUID + "." + str(series_number)
dataset.FrameOfReferenceUID = dataset.SeriesInstanceUID + ".0." + str(series_number)
dataset.PerformedProcedureStepID = "PERFSTEP" + series_suffix
dataset.RequestedProcedureID = "REQSTEP" + series_suffix
dataset.ScheduledProcedureStepID = "SCHEDSTEP" + series_suffix
dataset.SeriesDate = dataset.StudyDate
dataset.SeriesTime = datetime.time(patient_number, study_number, series_number).strftime("%H%M%S")
dataset.StationName = "STATIONNAME" + str(patient_number) + "." + str(study_number) + "." + str(series_number)
dataset.OperatorsName = "OPERATOR^FIRST^" + series_suffix
dataset.PerformingPhysicianName = "PERFORMING1^FIRST^" + series_suffix
request_attribute_item = pydicom.dataset.Dataset()
request_attribute_item.RequestedProcedureID = dataset.RequestedProcedureID
request_attribute_item.ScheduledProcedureStepID = dataset.ScheduledProcedureStepID
dataset.RequestAttributesSequence = pydicom.sequence.Sequence([request_attribute_item])
dataset.InstitutionName = "INSTITUTIONNAME" + series_suffix
dataset.InstitutionAddress = "INSTITUTIONADDRESS" + series_suffix
dataset.InstitutionalDepartmentName = "INSTITUTIONALDEPARTMENTNAME" + series_suffix
dataset.StationName = "STATIONNAME" + series_suffix
def _set_patient_attributes(dataset: pydicom.dataset.Dataset, patient_number: int) -> None:
dataset.PatientAddress = str(123 + patient_number) + " Fake Street"
dataset.PatientBirthDate = str(19830213 + patient_number)
dataset.PatientBirthTime = "13:14:0" + str(patient_number)
dataset.PatientID = "4MR" + str(patient_number)
dataset.PatientName = "CompressedSamples^MR" + str(patient_number)
dataset.OtherPatientIDs = "OTH" + dataset.PatientID
dataset.OtherPatientIDsSequence = [pydicom.dataset.Dataset()]
dataset.OtherPatientIDsSequence[0].PatientID = "OTHSEQ" + dataset.PatientID
dataset.OtherPatientNames = "Other" + str(dataset.PatientName)
dataset.PatientBirthName = "Birth" + str(dataset.PatientName)
dataset.PatientMotherBirthName = "Mother" + str(dataset.PatientName)
dataset.ResponsiblePerson = "Responsible" + str(dataset.PatientName)
def _update_patient_identity_removed(self, dataset: pydicom.dataset.Dataset) -> None:
if dataset.get("BurnedInAnnotation", "YES") == "NO":
dataset.PatientIdentityRemoved = "YES"
def _should_assume_annotation(self,
dataset: pydicom.dataset.Dataset) -> bool:
burned_in_annotation = dataset.get("BurnedInAnnotation")
return (self.assume_if == "if-yes" and burned_in_annotation
== "YES") or (self.assume_if == "unless-no"
and burned_in_annotation != "NO")