biomedisa 2024.5.18__py3-none-any.whl → 2024.5.20__py3-none-any.whl

biomedisa/__init__.py

@@ -1,15 +1,11 @@
 import os
-import sys
 import subprocess
 
 # from source base directory
 BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
 
-# pip base directory
-if not os.path.exists(os.path.join(BASE_DIR,'biomedisa_features')):
-    BASE_DIR = os.path.dirname(os.path.abspath(__file__))
-    # add BASE_DIR to PYTHONPATH for absolute imports
-    sys.path.append(BASE_DIR)
+# pip installation
+if not os.path.exists(os.path.join(BASE_DIR,'biomedisa/settings.py')):
 
     # metadata
     import importlib_metadata

biomedisa/deeplearning.py

@@ -1,7 +1,7 @@
 #!/usr/bin/python3
 ##########################################################################
 ##                                                                      ##
-##  Copyright (c) 2024 Philipp Lösel. All rights reserved.              ##
+##  Copyright (c) 2019-2024 Philipp Lösel. All rights reserved.         ##
 ##                                                                      ##
 ##  This file is part of the open source project biomedisa.             ##
 ##                                                                      ##
@@ -27,13 +27,11 @@
 ##                                                                      ##
 ##########################################################################
 
-import sys, os
-BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
-sys.path.append(BASE_DIR)
+import os
 import biomedisa
-import biomedisa_features.crop_helper as ch
-from biomedisa_features.keras_helper import *
-from biomedisa_features.biomedisa_helper import _error_, unique_file_path
+import biomedisa.features.crop_helper as ch
+from biomedisa.features.keras_helper import *
+from biomedisa.features.biomedisa_helper import _error_, unique_file_path
 from tensorflow.python.framework.errors_impl import ResourceExhaustedError
 import tensorflow as tf
 import numpy as np
@@ -43,6 +41,7 @@ import h5py
 import time
 import subprocess
 import glob
+import tempfile
 
 class Biomedisa(object):
      pass
@@ -120,7 +119,7 @@ def deep_learning(img_data, label_data=None, val_img_data=None, val_label_data=N
             project = os.path.splitext(bm.shortfilename)[0]
 
             # path to model
-            bm.path_to_model = BASE_DIR + f'/private_storage/images/{bm.username}/{project}.h5'
+            bm.path_to_model = biomedisa.BASE_DIR + f'/private_storage/images/{bm.username}/{project}.h5'
             if not bm.remote:
                 bm.path_to_model = unique_file_path(bm.path_to_model)
 
@@ -128,7 +127,7 @@ def deep_learning(img_data, label_data=None, val_img_data=None, val_label_data=N
             project = os.path.splitext(os.path.basename(bm.path_to_model))[0]
 
         # create pid object
-        from biomedisa_features.django_env import create_pid_object
+        from biomedisa.features.django_env import create_pid_object
         create_pid_object(os.getpid(), bm.remote, bm.queue, bm.img_id, (bm.path_to_model if bm.train else ''))
 
         # write in log file
@@ -227,81 +226,83 @@ def deep_learning(img_data, label_data=None, val_img_data=None, val_label_data=N
         crop_data = True if 'cropping_weights' in hf else False
         hf.close()
 
-        # extract image files
-        bm.tmp_dir = None
-        if bm.path_to_images is not None and (os.path.splitext(bm.path_to_images)[1]=='.tar' or bm.path_to_images[-7:]=='.tar.gz'):
-            path_to_dir = os.path.dirname(bm.path_to_images) + '/final.'+os.path.basename(bm.path_to_images)
-            if path_to_dir[-3:]=='.gz':
-                path_to_dir = path_to_dir[:-3]
-            if bm.django_env and not bm.remote:
-                path_to_dir = unique_file_path(path_to_dir)
-            bm.tmp_dir = BASE_DIR + '/tmp/' + id_generator(40)
-            tar = tarfile.open(bm.path_to_images)
-            tar.extractall(path=bm.tmp_dir)
-            tar.close()
-            bm.path_to_images = bm.tmp_dir
-            bm.save_cropped, bm.acwe = False, False
-            bm.clean, bm.fill = None, None
-
-        # list of images
-        path_to_finals = []
-        if bm.path_to_images is not None and os.path.isdir(bm.path_to_images):
-            files = glob.glob(bm.path_to_images+'/**/*', recursive=True)
-            bm.path_to_images = [f for f in files if os.path.isfile(f)]
-        else:
-            bm.path_to_images = [bm.path_to_images]
-
-        # loop over all images
-        for bm.path_to_image in bm.path_to_images:
-
-            # create path_to_final
-            if bm.path_to_image:
-                filename = os.path.basename(bm.path_to_image)
-                filename = os.path.splitext(filename)[0]
-                if filename[-4:] == '.nii':
-                    filename = filename[:-4]
-                bm.path_to_cropped_image = os.path.dirname(bm.path_to_image) + '/' + filename + '.cropped.tif'
-                if bm.django_env and not bm.remote and not bm.tmp_dir:
-                    bm.path_to_cropped_image = unique_file_path(bm.path_to_cropped_image)
-                filename = 'final.' + filename
-                bm.path_to_final = os.path.dirname(bm.path_to_image) + '/' + filename + extension
-                if bm.django_env and not bm.remote and not bm.tmp_dir:
-                    bm.path_to_final = unique_file_path(bm.path_to_final)
-
-            # crop data
-            region_of_interest, cropped_volume = None, None
-            if crop_data:
-                region_of_interest, cropped_volume = ch.crop_data(bm.path_to_image, bm.path_to_model, bm.path_to_cropped_image,
-                    bm.batch_size, bm.debug_cropping, bm.save_cropped, img_data, bm.x_range, bm.y_range, bm.z_range)
-
-            # load prediction data
-            img, img_header, z_shape, y_shape, x_shape, region_of_interest, img_data = load_prediction_data(bm.path_to_image,
-                channels, bm.x_scale, bm.y_scale, bm.z_scale, bm.no_scaling, normalize, normalization_parameters,
-                region_of_interest, img_data, img_header)
-
-            # make prediction
-            results, bm = predict_semantic_segmentation(bm, img, bm.path_to_model,
-                bm.z_patch, bm.y_patch, bm.x_patch, z_shape, y_shape, x_shape, bm.compression, header,
-                img_header, bm.stride_size, allLabels, bm.batch_size, region_of_interest,
-                bm.no_scaling, extension, img_data)
-
-            # results
-            if cropped_volume is not None:
-                results['cropped_volume'] = cropped_volume
-
-            # path to results
-            if bm.path_to_image:
-                path_to_finals.append(bm.path_to_final)
-
-        # write tar file and delete extracted image files
-        if bm.tmp_dir is not None and os.path.exists(bm.tmp_dir):
-            with tarfile.open(path_to_dir, 'w') as tar:
-                for file_path in path_to_finals:
-                    file_name = os.path.basename(file_path)
-                    tar.add(file_path, arcname=file_name)
-            shutil.rmtree(bm.tmp_dir)
-            bm.path_to_final = path_to_dir
-            bm.path_to_cropped_image = None
+        # make temporary directory
+        with tempfile.TemporaryDirectory() as temp_dir:
+
+            # extract image files from tar file
+            bm.tarfile = False
+            if bm.path_to_images is not None and (os.path.splitext(bm.path_to_images)[1]=='.tar' or bm.path_to_images[-7:]=='.tar.gz'):
+                bm.tarfile = True
+                path_to_result = os.path.dirname(bm.path_to_images) + '/final.'+os.path.basename(bm.path_to_images)
+                if path_to_result[-3:]=='.gz':
+                    path_to_result = path_to_result[:-3]
+                if bm.django_env and not bm.remote:
+                    path_to_result = unique_file_path(path_to_result)
+                tar = tarfile.open(bm.path_to_images)
+                tar.extractall(path=temp_dir)
+                tar.close()
+                bm.path_to_images = temp_dir
+                bm.save_cropped, bm.acwe = False, False
+                bm.clean, bm.fill = None, None
+
+            # list of images
+            path_to_finals = []
+            if bm.path_to_images is not None and os.path.isdir(bm.path_to_images):
+                files = glob.glob(bm.path_to_images+'/**/*', recursive=True)
+                bm.path_to_images = [f for f in files if os.path.isfile(f)]
+            else:
+                bm.path_to_images = [bm.path_to_images]
+
+            # loop over all images
+            for bm.path_to_image in bm.path_to_images:
+
+                # create path_to_final
+                if bm.path_to_image:
+                    filename = os.path.basename(bm.path_to_image)
+                    filename = os.path.splitext(filename)[0]
+                    if filename[-4:] == '.nii':
+                        filename = filename[:-4]
+                    bm.path_to_cropped_image = os.path.dirname(bm.path_to_image) + '/' + filename + '.cropped.tif'
+                    if bm.django_env and not bm.remote and not bm.tarfile:
+                        bm.path_to_cropped_image = unique_file_path(bm.path_to_cropped_image)
+                    filename = 'final.' + filename
+                    bm.path_to_final = os.path.dirname(bm.path_to_image) + '/' + filename + extension
+                    if bm.django_env and not bm.remote and not bm.tarfile:
+                        bm.path_to_final = unique_file_path(bm.path_to_final)
+
+                # crop data
+                region_of_interest, cropped_volume = None, None
+                if crop_data:
+                    region_of_interest, cropped_volume = ch.crop_data(bm.path_to_image, bm.path_to_model, bm.path_to_cropped_image,
+                        bm.batch_size, bm.debug_cropping, bm.save_cropped, img_data, bm.x_range, bm.y_range, bm.z_range)
+
+                # load prediction data
+                img, img_header, z_shape, y_shape, x_shape, region_of_interest, img_data = load_prediction_data(bm.path_to_image,
+                    channels, bm.x_scale, bm.y_scale, bm.z_scale, bm.no_scaling, normalize, normalization_parameters,
+                    region_of_interest, img_data, img_header)
+
+                # make prediction
+                results, bm = predict_semantic_segmentation(bm, img, bm.path_to_model,
+                    bm.z_patch, bm.y_patch, bm.x_patch, z_shape, y_shape, x_shape, bm.compression, header,
+                    img_header, bm.stride_size, allLabels, bm.batch_size, region_of_interest,
+                    bm.no_scaling, extension, img_data)
+
+                # results
+                if cropped_volume is not None:
+                    results['cropped_volume'] = cropped_volume
+
+                # path to results
+                if bm.path_to_image:
+                    path_to_finals.append(bm.path_to_final)
+
+            # write tar file and delete extracted image files
+            if bm.tarfile and os.path.exists(temp_dir):
+                with tarfile.open(path_to_result, 'w') as tar:
+                    for file_path in path_to_finals:
+                        file_name = os.path.basename(file_path)
+                        tar.add(file_path, arcname=file_name)
+                bm.path_to_final = path_to_result
+                bm.path_to_cropped_image = None
 
     # computation time
     t = int(time.time() - TIC)
@@ -316,7 +317,7 @@ def deep_learning(img_data, label_data=None, val_img_data=None, val_label_data=N
     # django environment
     if bm.django_env:
         from biomedisa_app.config import config
-        from biomedisa_features.django_env import post_processing
+        from biomedisa.features.django_env import post_processing
         validation=True if bm.validation_split or (bm.val_images is not None and bm.val_images[0] is not None) else False
         post_processing(bm.path_to_final, time_str, config['SERVER_ALIAS'], bm.remote, bm.queue,
             img_id=bm.img_id, label_id=bm.label_id, path_to_model=bm.path_to_model,
@@ -324,7 +325,7 @@ def deep_learning(img_data, label_data=None, val_img_data=None, val_label_data=N
             train=bm.train, predict=bm.predict, validation=validation)
 
         # write in log file
-        path_to_time = BASE_DIR + '/log/time.txt'
+        path_to_time = biomedisa.BASE_DIR + '/log/time.txt'
         with open(path_to_time, 'a') as timefile:
             if predict:
                 message = 'Successfully segmented ' + bm.shortfilename
@@ -487,7 +488,7 @@ if __name__ == '__main__':
             reference_image_path = bm.path_to_images
         bm.username = os.path.basename(os.path.dirname(reference_image_path))
         bm.shortfilename = os.path.basename(reference_image_path)
-        bm.path_to_logfile = BASE_DIR + '/log/logfile.txt'
+        bm.path_to_logfile = biomedisa.BASE_DIR + '/log/logfile.txt'
     else:
         bm.django_env = False
 
@@ -497,13 +498,9 @@ if __name__ == '__main__':
     try:
         deep_learning(None, **kwargs)
     except InputError:
-        if any(InputError.img_names):
-            remove_extracted_data(InputError.img_names, InputError.label_names)
         print(traceback.format_exc())
         bm = _error_(bm, f'{InputError.message}')
     except ch.InputError:
-        if any(ch.InputError.img_names):
-            remove_extracted_data(ch.InputError.img_names, ch.InputError.label_names)
         print(traceback.format_exc())
         bm = _error_(bm, f'{ch.InputError.message}')
     except MemoryError:

biomedisa/{biomedisa_features → features}/DataGenerator.py

@@ -1,6 +1,6 @@
 ##########################################################################
 ##                                                                      ##
-##  Copyright (c) 2024 Philipp Lösel. All rights reserved.              ##
+##  Copyright (c) 2019-2024 Philipp Lösel. All rights reserved.         ##
 ##                                                                      ##
 ##  This file is part of the open source project biomedisa.             ##
 ##                                                                      ##

biomedisa/{biomedisa_features → features}/DataGeneratorCrop.py

@@ -1,6 +1,6 @@
 ##########################################################################
 ##                                                                      ##
-##  Copyright (c) 2024 Philipp Lösel. All rights reserved.              ##
+##  Copyright (c) 2019-2024 Philipp Lösel. All rights reserved.         ##
 ##                                                                      ##
 ##  This file is part of the open source project biomedisa.             ##
 ##                                                                      ##

biomedisa/{biomedisa_features → features}/PredictDataGenerator.py

@@ -1,6 +1,6 @@
 ##########################################################################
 ##                                                                      ##
-##  Copyright (c) 2024 Philipp Lösel. All rights reserved.              ##
+##  Copyright (c) 2019-2024 Philipp Lösel. All rights reserved.         ##
 ##                                                                      ##
 ##  This file is part of the open source project biomedisa.             ##
 ##                                                                      ##

biomedisa/{biomedisa_features → features}/PredictDataGeneratorCrop.py

@@ -1,6 +1,6 @@
 ##########################################################################
 ##                                                                      ##
-##  Copyright (c) 2022 Philipp Lösel. All rights reserved.              ##
+##  Copyright (c) 2019-2024 Philipp Lösel. All rights reserved.         ##
 ##                                                                      ##
 ##  This file is part of the open source project biomedisa.             ##
 ##                                                                      ##

biomedisa/{biomedisa_features → features}/active_contour.py

@@ -1,7 +1,7 @@
 #!/usr/bin/python3
 ##########################################################################
 ##                                                                      ##
-##  Copyright (c) 2024 Philipp Lösel. All rights reserved.              ##
+##  Copyright (c) 2019-2024 Philipp Lösel. All rights reserved.         ##
 ##                                                                      ##
 ##  This file is part of the open source project biomedisa.             ##
 ##                                                                      ##
@@ -27,13 +27,10 @@
 ##                                                                      ##
 ##########################################################################
 
-import sys, os
-BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
-if not BASE_DIR in sys.path:
-    sys.path.append(BASE_DIR)
+import os
 import biomedisa
-from biomedisa_features.curvop_numba import curvop, evolution
-from biomedisa_features.biomedisa_helper import (unique_file_path, load_data, save_data,
+from biomedisa.features.curvop_numba import curvop, evolution
+from biomedisa.features.biomedisa_helper import (unique_file_path, load_data, save_data,
     pre_processing, img_to_uint8, silent_remove)
 import numpy as np
 import numba
@@ -143,7 +140,7 @@ def activeContour(data, labelData, alpha=1.0, smooth=1, steps=3,
     if bm.django_env:
         bm.username = os.path.basename(os.path.dirname(bm.path_to_data))
         bm.shortfilename = os.path.basename(bm.path_to_data)
-        bm.path_to_logfile = BASE_DIR + '/log/logfile.txt'
+        bm.path_to_logfile = biomedisa.BASE_DIR + '/log/logfile.txt'
 
     # pre-processing
     bm = pre_processing(bm)
@@ -234,13 +231,13 @@ def refinement(bm):
 
 def post_processing(path_to_acwe, image_id=None, friend_id=None, simple=False, remote=False):
     if remote:
-        with open(BASE_DIR + '/log/config_4', 'w') as configfile:
+        with open(biomedisa.BASE_DIR + '/log/config_4', 'w') as configfile:
             print(path_to_acwe, 'phantom', file=configfile)
     else:
         import django
         django.setup()
         from biomedisa_app.models import Upload
-        from biomedisa_features.create_slices import create_slices
+        from biomedisa.features.create_slices import create_slices
         from redis import Redis
         from rq import Queue
 
@@ -300,7 +297,7 @@ def init_active_contour(image_id, friend_id, label_id, simple=False):
 
     # get host information
     host = ''
-    host_base = BASE_DIR
+    host_base = biomedisa.BASE_DIR
     subhost, qsub_pid = None, None
     if 'REMOTE_QUEUE_HOST' in config:
         host = config['REMOTE_QUEUE_HOST']
@@ -313,8 +310,8 @@ def init_active_contour(image_id, friend_id, label_id, simple=False):
         if host:
 
             # command
-            cmd = ['python3', host_base+'/biomedisa_features/active_contour.py']
-            cmd += [image.pic.path.replace(BASE_DIR,host_base), friend.pic.path.replace(BASE_DIR,host_base)]
+            cmd = ['python3', host_base+'/biomedisa/features/active_contour.py']
+            cmd += [image.pic.path.replace(biomedisa.BASE_DIR,host_base), friend.pic.path.replace(biomedisa.BASE_DIR,host_base)]
             cmd += [f'-iid={image.id}', f'-fid={friend.id}', '-r']
 
             # command (append only on demand)
@@ -338,8 +335,8 @@ def init_active_contour(image_id, friend_id, label_id, simple=False):
 
             # send data to host
             success=0
-            success+=send_data_to_host(image.pic.path, host+':'+image.pic.path.replace(BASE_DIR,host_base))
-            success+=send_data_to_host(friend.pic.path, host+':'+friend.pic.path.replace(BASE_DIR,host_base))
+            success+=send_data_to_host(image.pic.path, host+':'+image.pic.path.replace(biomedisa.BASE_DIR,host_base))
+            success+=send_data_to_host(friend.pic.path, host+':'+friend.pic.path.replace(biomedisa.BASE_DIR,host_base))
 
             if success==0:
 
@@ -355,14 +352,14 @@ def init_active_contour(image_id, friend_id, label_id, simple=False):
                 subprocess.Popen(cmd).wait()
 
                 # config
-                success = subprocess.Popen(['scp', host+':'+host_base+'/log/config_4', BASE_DIR+'/log/config_4']).wait()
+                success = subprocess.Popen(['scp', host+':'+host_base+'/log/config_4', biomedisa.BASE_DIR+'/log/config_4']).wait()
 
                 if success==0:
-                    with open(BASE_DIR + '/log/config_4', 'r') as configfile:
+                    with open(biomedisa.BASE_DIR + '/log/config_4', 'r') as configfile:
                         acwe_on_host, _ = configfile.read().split()
 
                     # local file names
-                    path_to_acwe = unique_file_path(acwe_on_host.replace(host_base,BASE_DIR))
+                    path_to_acwe = unique_file_path(acwe_on_host.replace(host_base,biomedisa.BASE_DIR))
 
                     # get results
                     subprocess.Popen(['scp', host+':'+acwe_on_host, path_to_acwe]).wait()

biomedisa/{biomedisa_features → features}/assd.py

@@ -1,6 +1,6 @@
 ##########################################################################
 ##                                                                      ##
-##  Copyright (c) 2024 Philipp Lösel. All rights reserved.              ##
+##  Copyright (c) 2019-2024 Philipp Lösel. All rights reserved.         ##
 ##                                                                      ##
 ##  This file is part of the open source project biomedisa.             ##
 ##                                                                      ##