Use canonicaljson instead of json.dumps(..., sort_keys=True) for …

…hashing
mlexchange · Mar 4, 2024 · f0d6a42 · f0d6a42
1 parent 212530d
commit f0d6a42
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 3 deletions.
diff --git a/requirements.txt b/requirements.txt
@@ -33,3 +33,4 @@ scipy
 dash-extensions==1.0.1
 dash-bootstrap-components==1.5.0
 dash_auth==2.0.0
+canonicaljson
diff --git a/utils/annotations.py b/utils/annotations.py
@@ -1,8 +1,8 @@
 import hashlib
 import io
-import json
 import zipfile
 
+import canonicaljson
 import numpy as np
 import scipy.sparse as sp
 from matplotlib.path import Path
@@ -62,8 +62,8 @@ def get_annotation_classes(self):
 
     def get_annotations_hash(self):
         hash_object = hashlib.md5()
-        hash_object.update(json.dumps(self.annotations, sort_keys=True).encode())
-        hash_object.update(json.dumps(self.annotation_classes, sort_keys=True).encode())
+        hash_object.update(canonicaljson.encode_canonical_json(self.annotations))
+        hash_object.update(canonicaljson.encode_canonical_json(self.annotation_classes))
         return hash_object.hexdigest()
 
     def get_annotation_mask_as_bytes(self):