fix download retry bug, add get_cache_dir to hut/util

modelscope/hub/file_download.py

@@ -224,10 +224,10 @@ def download_part_with_retry(params):
                 with open(part_file_name, 'rb') as f:
                     partial_length = f.seek(0, io.SEEK_END)
                     progress.update(partial_length)
-            start = start + partial_length
-            if start > end:
+            download_start = start + partial_length
+            if download_start > end:
                 break  # this part is download completed.
-            get_headers['Range'] = 'bytes=%s-%s' % (start, end)
+            get_headers['Range'] = 'bytes=%s-%s' % (download_start, end)
             with open(part_file_name, 'ab+') as f:
                 r = requests.get(
                     url,

modelscope/hub/utils/utils.py

@@ -3,6 +3,7 @@
 import hashlib
 import os
 from datetime import datetime
+from pathlib import Path
 from typing import Optional
 
 import requests
@@ -28,6 +29,21 @@ def model_id_to_group_owner_name(model_id):
     return group_or_owner, name
 
 
+def get_cache_dir(model_id: Optional[str] = None):
+    """cache dir precedence:
+        function parameter > environment > ~/.cache/modelscope/hub
+    Args:
+        model_id (str, optional): The model id.
+    Returns:
+        str: the model_id dir if model_id not None, otherwise cache root dir.
+    """
+    default_cache_dir = Path.home().joinpath('.cache', 'modelscope')
+    base_path = os.getenv('MODELSCOPE_CACHE',
+                          os.path.join(default_cache_dir, 'hub'))
+    return base_path if model_id is None else os.path.join(
+        base_path, model_id + '/')
+
+
 def get_release_datetime():
     if MODELSCOPE_SDK_DEBUG in os.environ:
         rt = int(round(datetime.now().timestamp()))