def __segment_nifti(self, source_file, filepath_only=False):
# assert __is_nifti(source_file)
payload = {'source_file': source_file}
print("filepath only ", filepath_only)
response_dict = self.container_requester.send_request_to_worker(payload,
self.worker_hostname,
self.worker_port,
self.segment_muscle_request_name)
rel_seg_path = response_dict["segmentation"]
data_share = os.environ["DATA_SHARE_PATH"]
segmentation_path = os.path.join(data_share, rel_seg_path)
abs_source_file = os.path.join(data_share, source_file)
if filepath_only:
return segmentation_path, abs_source_file
print("reading muscle segmentation from", segmentation_path)
segmentation = nifti_reader.read_nifti_image(segmentation_path)
original = nifti_reader.read_nifti_image(abs_source_file)
return segmentation, original
def read_image(source_dir, filepath_only=False):
data_share = os.environ["DATA_SHARE_PATH"]
full_source = os.path.join(data_share, source_dir)
if filepath_only:
return full_source
image = read_nifti_image(full_source)
return image
def convert(source_dir, filepath_only=False):
data_share = os.environ["DATA_SHARE_PATH"]
converter = Converter(ContainerRequester())
nifti_filename = converter.convert_dcm_to_nifti(source_dir)
nifti_filename = os.path.join(data_share, nifti_filename)
if filepath_only:
return nifti_filename
return read_nifti_image(nifti_filename)
def __generate_maps(self, source_file, filepath_only):
payload = {"source_file": source_file}
response_dict = self.container_requester.send_request_to_worker(
payload, self.worker_hostname, self.worker_port, self.request_name)
rel_mask_volume_path = response_dict["auxiliary_volume"]
rel_detection_volume_path = response_dict["detection_volume"]
data_share = os.environ["DATA_SHARE_PATH"]
mask_volume_path = os.path.join(data_share, rel_mask_volume_path)
detection_volume_path = os.path.join(data_share,
rel_detection_volume_path)
if filepath_only:
return mask_volume_path, detection_volume_path
mask_volume = nifti_reader.read_nifti_image(mask_volume_path)
detection_volume = nifti_reader.read_nifti_image(detection_volume_path)
return mask_volume, detection_volume
def __docker_segment(self, source_dir, model_name, filepath_only):
payload = {'source_dir': source_dir, 'model_name': model_name}
response_dict = self.container_requester.send_request_to_worker(
payload, self.worker_hostname, self.worker_port,
self.segment_request_name)
rel_seg_path = response_dict["segmentation"]
data_share = os.environ["DATA_SHARE_PATH"]
segmentation_path = os.path.join(data_share, rel_seg_path)
if filepath_only:
return segmentation_path
segmentation = read_nifti_image(segmentation_path)
return segmentation
def display_volume_and_slice_information(self,
input_nifti_path,
lung_seg_path=None,
muscle_seg=None,
lesion_detection=None,
lesion_attention=None,
lesion_detection_seg=None,
lesion_mask_seg=None,
fat_report=None,
fat_interval=None):
original_array = self.__get_array_from_image_path(input_nifti_path)
# lung_seg_path may be None, in which case fat_report_displayer doesn't display anything
lungmask_displayer \
= LungmaskSegmentationDisplayer(input_nifti_path, lung_seg_path, streamlit_wrapper=self.st,
download_displayer=self.download_class(streamlit_wrapper=self.st))
# lung_seg is None if lung_seg_path is None
lung_seg = lungmask_displayer.get_seg_array()
# fat_report may be None, in which case fat_report_displayer doesn't display anything
fat_report_displayer = FatReportDisplayer(
original_array,
fat_report,
fat_interval=fat_interval,
streamlit_wrapper=self.st,
download_displayer=self.download_class(streamlit_wrapper=self.st))
lungmask_displayer.download_button()
fat_report_displayer.download_button()
# may be None
fat_report_cm3 = fat_report_displayer.get_converted_report()
input_download_displayer = self.download_class(
streamlit_wrapper=self.st)
input_download_displayer.display(
os.path.split(input_nifti_path)[1], "Input Image")
detection_array = None
attention_array = None
muscle_seg_array = None
detection_seg_array = None
mask_seg_array = None
if lesion_detection is not None:
detection_array = read_nifti_image(lesion_detection)
detection_array = sitk.GetArrayFromImage(detection_array)
detection_download_displayer = self.download_class(
streamlit_wrapper=self.st)
detection_download_displayer.display(
os.path.split(lesion_detection)[1], "Lesion Detection Volume")
if lesion_attention is not None:
attention_array = read_nifti_image(lesion_attention)
attention_array = sitk.GetArrayFromImage(attention_array)
attention_download_displayer = self.download_class(
streamlit_wrapper=self.st)
attention_download_displayer.display(
os.path.split(lesion_attention)[1], "Attention Volume")
if muscle_seg is not None:
muscle_seg_array = read_nifti_image(muscle_seg)
muscle_seg_array = sitk.GetArrayFromImage(muscle_seg_array)
muscle_download_displayer = self.download_class(
streamlit_wrapper=self.st)
muscle_download_displayer.display(
os.path.split(muscle_seg)[1], "Muscle Segmentation")
if lesion_detection_seg is not None:
detection_seg_array = read_nifti_image(lesion_detection_seg)
detection_seg_array = sitk.GetArrayFromImage(detection_seg_array)
detection_seg_download_displayer = self.download_class(
streamlit_wrapper=self.st)
detection_seg_download_displayer.display(
os.path.split(lesion_detection_seg)[1],
"Lesion Detection Segmentation")
if lesion_mask_seg is not None:
mask_seg_array = read_nifti_image(lesion_mask_seg)
mask_seg_array = sitk.GetArrayFromImage(mask_seg_array)
mask_seg_download_displayer = self.download_class(
streamlit_wrapper=self.st)
mask_seg_download_displayer.display(
os.path.split(lesion_mask_seg)[1], "Lesion Detection Mask")
lungmask_displayer.display()
fat_report_displayer.display()
self.__display_information_rows(original_array, lung_seg,
muscle_seg_array, detection_array,
attention_array, detection_seg_array,
mask_seg_array, fat_report_cm3)
def __get_array_from_image_path(self, path):
image = read_nifti_image(path)
return sitk.GetArrayFromImage(image)