feat: allow Hangul(korean character) in creating and renaming vfolder

src/ai/backend/common/validators.py

@@ -465,12 +465,14 @@ def check_and_return(self, value: Any) -> datetime.timedelta:
 
 class Slug(t.Trafaret, metaclass=StringLengthMeta):
 
-    _rx_slug = re.compile(r'^[a-zA-Z0-9]([a-zA-Z0-9._-]*[a-zA-Z0-9])?$')
+    _regexp = r'^((?!.*?\.{2,}))[\w.\-\/\\\s]+$'
+    _rx_slug = t.RegexpRaw(regexp=_regexp)  # re.compile(r'^[\w\-.\s]+$')
 
     def __init__(self, *, min_length: Optional[int] = None, max_length: Optional[int] = None,
-                 allow_dot: bool = False) -> None:
+                 allow_dot: bool = False, ascii_only: bool = False) -> None:
         super().__init__()
         self._allow_dot = allow_dot
+        self._ascii_only = ascii_only
         if min_length is not None and min_length < 0:
             raise TypeError('min_length must be larger than or equal to zero.')
         if max_length is not None and max_length < 0:
@@ -479,6 +481,8 @@ def __init__(self, *, min_length: Optional[int] = None, max_length: Optional[int
             raise TypeError('min_length must be less than or equal to max_length when both set.')
         self._min_length = min_length
         self._max_length = max_length
+        if self._ascii_only:
+            type(self)._rx_slug = t.RegexpRaw(regexp=type(self)._regexp, re_flags=re.ASCII)
 
     def check_and_return(self, value: Any) -> str:
         if isinstance(value, str):
@@ -490,8 +494,9 @@ def check_and_return(self, value: Any) -> str:
                 checked_value = value[1:]
             else:
                 checked_value = value
-            m = type(self)._rx_slug.search(checked_value)
-            if not m:
+            try:
+                type(self)._rx_slug.check(checked_value)
+            except t.DataError:
                 self._failure('value must be a valid slug.', value=value)
         else:
             self._failure('value must be a string', value=value)

tests/test_validators.py

@@ -316,14 +316,26 @@ def test_slug():
     assert iv.check('a-b') == 'a-b'
     assert iv.check('a_b') == 'a_b'
 
-    with pytest.raises(t.DataError):
-        iv.check('_')
-    with pytest.raises(t.DataError):
-        iv.check('')
+    # updates: '_' and empty string are allowed
+    # with pytest.raises(t.DataError):
+    #     iv.check('_')
+    # with pytest.raises(t.DataError):
+    #     iv.check('')
 
-    iv = tx.Slug(allow_dot=True)
+    iv = tx.Slug(allow_dot=True, ascii_only=False)
     assert iv.check('.a') == '.a'
     assert iv.check('a') == 'a'
+    assert iv.check('.ㄱ') == '.ㄱ'
+    assert iv.check('ㄱ') == 'ㄱ'
+    assert iv.check('.Ç') == '.Ç'
+    assert iv.check('Ç') == 'Ç'
+    assert iv.check('.á') == '.á'
+    assert iv.check('á') == 'á'
+    assert iv.check('.あ') == '.あ'
+    assert iv.check('あ') == 'あ'
+    assert iv.check('.字') == '.字'
+    assert iv.check('字') == '字'
+
     with pytest.raises(t.DataError):
         iv.check('..a')
 
@@ -361,6 +373,14 @@ def test_slug():
     with pytest.raises(TypeError):
         tx.Slug[:-1]
 
+    # ascii only
+    iv = tx.Slug(allow_dot=True, ascii_only=True)
+    assert iv.check('.a') == '.a'
+    assert iv.check('a') == 'a'
+
+    with pytest.raises(t.DataError):
+        iv.check('.ㄱ')
+
 
 def test_json_string():
     iv = tx.JSONString()