qmenta-anon 2.2.1.dev385__tar.gz → 2.2.2.dev386__tar.gz

{qmenta_anon-2.2.1.dev385 → qmenta_anon-2.2.2.dev386}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: qmenta-anon
-Version: 2.2.1.dev385
+Version: 2.2.2.dev386
 Summary: The qmenta-anon library is used to anonymize files before transferring them to the QMENTA platform.
 Author: QMENTA
 Author-email: dev@qmenta.com

{qmenta_anon-2.2.1.dev385 → qmenta_anon-2.2.2.dev386}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "qmenta-anon"
-version = "2.2.1.dev385"
+version = "2.2.2.dev386"
 description = "The qmenta-anon library is used to anonymize files before transferring them to the QMENTA platform."
 authors = ["QMENTA <dev@qmenta.com>"]
 classifiers = [

{qmenta_anon-2.2.1.dev385 → qmenta_anon-2.2.2.dev386}/src/qmenta/anon/dicom.py

@@ -330,13 +330,53 @@ def anonymise_dicom_dataset(dcm, actions=None):
 
 
 def anonymise(filename, actions=None):
-    header = pydicom.read_file(filename, force=True)
+    try:
+        header = pydicom.read_file(filename)
+        invalid_dicom = False
+    except pydicom.errors.InvalidDicomError:
+        # 1. Read the legacy file
+        header = pydicom.read_file(filename, force=True)
+        invalid_dicom = True
+
+        # 2. Assess and safely default endianness and VR
+        if not hasattr(header, 'is_little_endian') or header.is_little_endian is None:
+            header.is_little_endian = True
+
+        if not hasattr(header, 'is_implicit_VR') or header.is_implicit_VR is None:
+            header.is_implicit_VR = True
+
+        # 3. Create the modern file meta header
+        if not hasattr(header, 'file_meta') or not header.file_meta:
+            header.file_meta = pydicom.dataset.FileMetaDataset()
+
+        # Dynamically assign the correct TransferSyntaxUID based on the dataset's properties
+        if header.is_little_endian and header.is_implicit_VR:
+            header.file_meta.TransferSyntaxUID = pydicom.uid.ImplicitVRLittleEndian
+        elif header.is_little_endian and not header.is_implicit_VR:
+            header.file_meta.TransferSyntaxUID = pydicom.uid.ExplicitVRLittleEndian
+        elif not header.is_little_endian and not header.is_implicit_VR:
+            header.file_meta.TransferSyntaxUID = pydicom.uid.ExplicitVRBigEndian
+        else:
+            # Fallback for the highly unusual Implicit VR Big Endian
+            header.file_meta.TransferSyntaxUID = pydicom.uid.ImplicitVRLittleEndian
+
+        # Copy the SOP UIDs to the header
+        if 'SOPClassUID' in header:
+            header.file_meta.MediaStorageSOPClassUID = header.SOPClassUID
+
+        if 'SOPInstanceUID' in header:
+            header.file_meta.MediaStorageSOPInstanceUID = header.SOPInstanceUID
+        else:
+            header.file_meta.MediaStorageSOPInstanceUID = pydicom.uid.generate_uid()
 
     actions = actions or DICOM_ANON_MIN_SUPP_55
     anonymise_dicom_dataset(header, actions)
     _updateMetaInfo(header)
 
-    header.save_as(filename)
+    if invalid_dicom:
+        header.save_as(filename, write_like_original=False)
+    else:
+        header.save_as(filename)
     return {"OK": True, "error_tags": []}
 
 
@@ -559,7 +599,45 @@ class RelativeTimeAnonymiser:
             when trying to anonymise multiple DICOM headers of which the
             datetimes to anonymise span more than 24h.
         """
-        header = pydicom.read_file(filename, force=True)
+        try:
+            header = pydicom.read_file(filename)
+            invalid_dicom = False
+        except pydicom.errors.InvalidDicomError:
+            # 1. Read the legacy file
+            header = pydicom.read_file(filename, force=True)
+            invalid_dicom = True
+
+            # 2. Assess and safely default endianness and VR
+            if not hasattr(header, 'is_little_endian') or header.is_little_endian is None:
+                header.is_little_endian = True
+
+            if not hasattr(header, 'is_implicit_VR') or header.is_implicit_VR is None:
+                header.is_implicit_VR = True
+
+            # 3. Create the modern file meta header
+            if not hasattr(header, 'file_meta') or not header.file_meta:
+                header.file_meta = pydicom.dataset.FileMetaDataset()
+
+            # Dynamically assign the correct TransferSyntaxUID based on the dataset's properties
+            if header.is_little_endian and header.is_implicit_VR:
+                header.file_meta.TransferSyntaxUID = pydicom.uid.ImplicitVRLittleEndian
+            elif header.is_little_endian and not header.is_implicit_VR:
+                header.file_meta.TransferSyntaxUID = pydicom.uid.ExplicitVRLittleEndian
+            elif not header.is_little_endian and not header.is_implicit_VR:
+                header.file_meta.TransferSyntaxUID = pydicom.uid.ExplicitVRBigEndian
+            else:
+                # Fallback for the highly unusual Implicit VR Big Endian
+                header.file_meta.TransferSyntaxUID = pydicom.uid.ImplicitVRLittleEndian
+
+            # Copy the SOP UIDs to the header
+            if 'SOPClassUID' in header:
+                header.file_meta.MediaStorageSOPClassUID = header.SOPClassUID
+
+            if 'SOPInstanceUID' in header:
+                header.file_meta.MediaStorageSOPInstanceUID = header.SOPInstanceUID
+            else:
+                header.file_meta.MediaStorageSOPInstanceUID = pydicom.uid.generate_uid()
+
         self.anonymise_datetime(header)
 
         actions = actions or DICOM_ANON_MIN_SUPP_55
@@ -572,4 +650,7 @@ class RelativeTimeAnonymiser:
         anonymise_dicom_dataset(header, restricted_actions)
         _updateMetaInfo(header)
 
-        header.save_as(filename)
+        if invalid_dicom:
+            header.save_as(filename, write_like_original=False)
+        else:
+            header.save_as(filename)