improve image hashing

generic/models.py

@@ -200,7 +200,18 @@ class SeriesImageBase(models.Model):
         null=True,
     )
 
+    # We store two hashes of the dicom data to uniquely identify the file
+    # and allow duplicate detection
+
+    # The first is a md5 hash of the pixel array data
+    # this is used because I can't get the direct pixel data to match from cornerstone
+    # (and struggle to get blake3 working in the browser)
     image_md5_hash = models.CharField(max_length=32, null=True, blank=True)
+
+    # The second is a blake3 hash of the direct pixel data
+    # This is much faster when using pydicom
+    image_blake3_hash = models.CharField(max_length=64, null=True, blank=True)
+
     is_dicom = models.BooleanField(default=False)
 
     removed = models.BooleanField(
@@ -231,14 +242,18 @@ class SeriesImageBase(models.Model):
         """Override save method to add image hash"""
         if self.image:
             if not self.image_md5_hash:
-                image_hash, is_dicom = get_image_hash(self.image)
+                image_hash, is_dicom = get_image_hash(self.image, hash_type="md5", direct_pixel_data=False)
 
                 self.is_dicom = is_dicom
                 self.image_md5_hash = image_hash
 
-                # Hack for tests
-                if image_hash != "12345ABCD":
-                    super().save(*args, **kwargs)  # Call the "real" save() method.
+            if not self.image_blake3_hash:
+                image_blake3_hash, is_dicom = get_image_hash(self.image, hash_type="blake3", direct_pixel_data=True)
+                self.image_blake3_hash = image_blake3_hash
+
+                ## Hack for tests
+                #if image_hash != "12345ABCD":
+                #    super().save(*args, **kwargs)  # Call the "real" save() method.
 
         super().save(*args, **kwargs)  # Call the "real" save() method.