From 288763037d32bc54fc5c3a0be5258528eba69dff Mon Sep 17 00:00:00 2001
From: Jakub Tymejczyk <jakub@tymejczyk.pl>
Date: Wed, 23 Aug 2023 13:46:25 +0200
Subject: [PATCH] normalize filenames

---
 src/sync_drive.py  | 13 ++++++++++---
 src/sync_photos.py |  8 +++++++-
 2 files changed, 17 insertions(+), 4 deletions(-)

diff --git a/src/sync_drive.py b/src/sync_drive.py
index 4a8a978c7..aa975d665 100644
--- a/src/sync_drive.py
+++ b/src/sync_drive.py
@@ -5,6 +5,7 @@
 import os
 import re
 import time
+import unicodedata
 from pathlib import Path
 from shutil import copyfileobj, rmtree, unpack_archive
 
@@ -78,12 +79,13 @@ def process_folder(item, destination_path, filters, ignore, root):
     if not (item and destination_path and root):
         return None
     new_directory = os.path.join(destination_path, item.name)
+    new_directory_norm = unicodedata.normalize('NFC', new_directory)
     if not wanted_folder(
-        filters=filters, ignore=ignore, folder_path=new_directory, root=root
+        filters=filters, ignore=ignore, folder_path=new_directory_norm, root=root
     ):
         LOGGER.debug(f"Skipping the unwanted folder {new_directory} ...")
         return None
-    os.makedirs(new_directory, exist_ok=True)
+    os.makedirs(new_directory_norm, exist_ok=True)
     return new_directory
 
 
@@ -152,6 +154,10 @@ def process_package(local_file):
         os.rename(local_file, archive_file)
         LOGGER.info(f"Unpacking {archive_file} to {os.path.dirname(archive_file)}")
         unpack_archive(filename=archive_file, extract_dir=os.path.dirname(archive_file))
+        try:
+            os.rename(unicodedata.normalize('NFD', local_file), local_file)
+        except:
+            LOGGER.warning("Normalizing failed")
         os.remove(archive_file)
     elif "application/gzip" == magic_object.from_file(filename=local_file):
         archive_file += ".gz"
@@ -203,6 +209,7 @@ def process_file(item, destination_path, filters, ignore, files):
     if not (item and destination_path and files is not None):
         return False
     local_file = os.path.join(destination_path, item.name)
+    local_file = unicodedata.normalize('NFC', local_file)
     if not wanted_file(filters=filters, ignore=ignore, file_path=local_file):
         return False
     files.add(local_file)
@@ -267,7 +274,7 @@ def sync_directory(
                 if not new_folder:
                     continue
                 try:
-                    files.add(new_folder)
+                    files.add(unicodedata.normalize('NFC', new_folder))
                     files.update(
                         sync_directory(
                             drive=item,
diff --git a/src/sync_photos.py b/src/sync_photos.py
index bc27f8eb1..f536bfa87 100644
--- a/src/sync_photos.py
+++ b/src/sync_photos.py
@@ -4,6 +4,7 @@
 import os
 import shutil
 import time
+import unicodedata
 from pathlib import Path
 
 from icloudpy import exceptions
@@ -33,11 +34,16 @@ def generate_file_name(photo, file_size, destination_path):
         destination_path,
         f'{"__".join([name, file_size, base64.urlsafe_b64encode(photo.id.encode()).decode()])}.{extension}',
     )
+
+    file_size_id_path_norm = unicodedata.normalize('NFC', file_size_id_path)
+
     if os.path.isfile(file_path):
         os.rename(file_path, file_size_id_path)
     if os.path.isfile(file_size_path):
         os.rename(file_size_path, file_size_id_path)
-    return file_size_id_path
+    if os.path.isfile(file_size_id_path):
+        os.rename(file_size_id_path, file_size_id_path_norm)
+    return file_size_id_path_norm
 
 
 def photo_exists(photo, file_size, local_path):