fix: setup hashlib for fips compliance (#123)

Signed-off-by: Michele Dolfi <[email protected]>

Author: dolfim-ibm

docling_parse/pdf_parser.py

@@ -455,7 +455,7 @@ def load(
             success = self._load_document(key=key, filename=str(path_or_stream))
 
         elif isinstance(path_or_stream, BytesIO):
-            hasher = hashlib.md5()
+            hasher = hashlib.sha256(usedforsecurity=False)
 
             while chunk := path_or_stream.read(8192):
                 hasher.update(chunk)

docling_parse/processing_dir.py

@@ -73,7 +73,7 @@ def fetch_files_from_disk(directory, recursive, task_queue):
 
         file_name = str(Path(filename).resolve())
 
-        hash_object = hashlib.sha256(filename.encode())
+        hash_object = hashlib.sha256(filename.encode(), usedforsecurity=False)
         file_hash = hash_object.hexdigest()
 
         # Create a FileTask object

docling_parse/processing_s3.py

@@ -171,7 +171,7 @@ def fetch_files_from_s3(
                 data = response["Body"].read()
 
                 # Generate a hash (e.g., SHA-256)
-                hash_object = hashlib.sha256(data)
+                hash_object = hashlib.sha256(data, usedforsecurity=False)
                 file_hash = hash_object.hexdigest()
 
                 local_file = os.path.join(local_dir, f"{file_hash}.pdf")
@@ -248,7 +248,7 @@ def retrieve_file(s3_client, task):
     task.data = BytesIO(response["Body"].read())  # response["Body"].read()
 
     # Generate a hash (e.g., SHA-256)
-    hash_object = hashlib.sha256(task.data.read())
+    hash_object = hashlib.sha256(task.data.read(), usedforsecurity=False)
     task.file_hash = str(hash_object.hexdigest())
 
     return task