scriptconv 0.0.3a14__tar.gz → 0.0.3a15__tar.gz

{scriptconv-0.0.3a14 → scriptconv-0.0.3a15}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scriptconv
-Version: 0.0.3a14
+Version: 0.0.3a15
 Summary: Zero-dependency script & phoneme-notation core — ISO-15924 detection & metadata, IPA↔ARPABET/X-SAMPA/Lexique/Kirshenbaum/Cotovía/RFE, Buckwalter↔Arabic, Hangul→jamo, kana
 License: Apache-2.0
 Project-URL: Homepage, https://github.com/TigreGotico/scriptconv

{scriptconv-0.0.3a14 → scriptconv-0.0.3a15}/scriptconv/phonemizers/base.py

@@ -45,8 +45,9 @@ class BasePhonemizer(metaclass=abc.ABCMeta):
         # optional (text, lang) -> str hook run before chunking; see module
         # docstring — scriptconv performs no normalization of its own
         self.normalizer = normalizer
-        # local path to a phonikud ONNX model (Hebrew diacritization);
-        # scriptconv never downloads — the consumer resolves the file
+        # local path to a phonikud ONNX model (Hebrew diacritization), or a
+        # zero-arg callable resolving one lazily; scriptconv never downloads
+        # — the consumer resolves the file
         self.phonikud_model = phonikud_model
 
         # diacritizer model name, for languages that need one. Arabic uses
@@ -59,7 +60,9 @@ class BasePhonemizer(metaclass=abc.ABCMeta):
     @property
     def phonikud(self):
         if self._phonikud is None:
-            if not self.phonikud_model:
+            model = self.phonikud_model() if callable(self.phonikud_model) \
+                else self.phonikud_model
+            if not model:
                 raise ValueError(
                     "Hebrew diacritization needs a local phonikud ONNX model: "
                     "pass phonikud_model=<path> (scriptconv never downloads "
@@ -70,7 +73,7 @@ class BasePhonemizer(metaclass=abc.ABCMeta):
                 raise ImportError(
                     "Hebrew diacritization needs phonikud-onnx — install "
                     "with `pip install scriptconv[he]`") from None
-            self._phonikud = Phonikud(self.phonikud_model)
+            self._phonikud = Phonikud(model)
         return self._phonikud
 
     def tashkeel(self, model: Optional[str] = None):

{scriptconv-0.0.3a14 → scriptconv-0.0.3a15}/scriptconv/version.py

@@ -2,7 +2,7 @@
 VERSION_MAJOR = 0
 VERSION_MINOR = 0
 VERSION_BUILD = 3
-VERSION_ALPHA = 14
+VERSION_ALPHA = 15
 # END_VERSION_BLOCK
 
 VERSION_STR = f"{VERSION_MAJOR}.{VERSION_MINOR}.{VERSION_BUILD}"

{scriptconv-0.0.3a14 → scriptconv-0.0.3a15}/scriptconv.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scriptconv
-Version: 0.0.3a14
+Version: 0.0.3a15
 Summary: Zero-dependency script & phoneme-notation core — ISO-15924 detection & metadata, IPA↔ARPABET/X-SAMPA/Lexique/Kirshenbaum/Cotovía/RFE, Buckwalter↔Arabic, Hangul→jamo, kana
 License: Apache-2.0
 Project-URL: Homepage, https://github.com/TigreGotico/scriptconv

{scriptconv-0.0.3a14 → scriptconv-0.0.3a15}/tests/test_phonemizers_base.py

@@ -159,3 +159,18 @@ class TestFacadeAndGraph(unittest.TestCase):
         from scriptconv import phonemizers
         g = DEFAULT_GRAPH.extend(phonemizers.register)
         self.assertTrue(g.can_convert("text", "arpa"))
+
+
+class TestPhonikudModelResolver(unittest.TestCase):
+    def test_callable_resolver_invoked_lazily(self):
+        calls = []
+
+        def resolver():
+            calls.append(1)
+            return ""  # resolves to nothing -> still the explicit ValueError
+
+        g = GraphemePhonemizer(phonikud_model=resolver)
+        self.assertEqual(calls, [])  # not resolved at construction
+        with self.assertRaises(ValueError):
+            g.add_diacritics("שלום", "he")
+        self.assertEqual(calls, [1])