feat(omnivoice): add Vietnamese (vi) to the language list

app/src/lib/constants/languages.ts

@@ -32,6 +32,7 @@ export const ALL_LANGUAGES = {
   sv: 'Swedish',
   sw: 'Swahili',
   tr: 'Turkish',
+  vi: 'Vietnamese',
   zh: 'Chinese',
 } as const;
 
@@ -96,6 +97,7 @@ export const ENGINE_LANGUAGES: Record<string, readonly LanguageCode[]> = {
     'sv',
     'sw',
     'tr',
+    'vi',
     'zh',
   ],
 } as const;

backend/backends/__init__.py

@@ -468,6 +468,7 @@ def _get_non_qwen_tts_configs() -> list[ModelConfig]:
                 "sv",
                 "sw",
                 "tr",
+                "vi",
             ],
         ),
     ]

backend/backends/omnivoice_backend.py

@@ -21,7 +21,7 @@
     pitch, whisper, English accent, Chinese dialect), not the free-form prose
     Qwen CustomVoice accepts. Anything outside it raises upstream.
 
-Languages: the 22 codes Voicebox already knows that OmniVoice also covers,
+Languages: the 23 codes Voicebox already knows that OmniVoice also covers,
 plus Arabic via the ISO 639-3 code ``arb``. OmniVoice supports far more, but
 exposing them means widening the language enum across the API and frontend.
 """

backend/models.py

@@ -18,7 +18,7 @@ class VoiceProfileCreate(BaseModel):
     name: str = Field(..., min_length=1, max_length=100)
     description: Optional[str] = Field(None, max_length=500)
     language: str = Field(
-        default="en", pattern="^(zh|en|ja|ko|de|fr|ru|pt|es|it|he|ar|da|el|fi|hi|ms|nl|no|pl|sv|sw|tr)$"
+        default="en", pattern="^(zh|en|ja|ko|de|fr|ru|pt|es|it|he|ar|da|el|fi|hi|ms|nl|no|pl|sv|sw|tr|vi)$"
     )
     voice_type: Optional[str] = Field(default="cloned", pattern="^(cloned|preset|designed)$")
     preset_engine: Optional[str] = Field(None, max_length=50)
@@ -81,7 +81,7 @@ class GenerationRequest(BaseModel):
 
     profile_id: str
     text: str = Field(..., min_length=1, max_length=50000)
-    language: str = Field(default="en", pattern="^(zh|en|ja|ko|de|fr|ru|pt|es|it|he|ar|da|el|fi|hi|ms|nl|no|pl|sv|sw|tr)$")
+    language: str = Field(default="en", pattern="^(zh|en|ja|ko|de|fr|ru|pt|es|it|he|ar|da|el|fi|hi|ms|nl|no|pl|sv|sw|tr|vi)$")
     seed: Optional[int] = Field(None, ge=0)
     model_size: Optional[str] = Field(default="1.7B", pattern="^(1\\.7B|0\\.6B|1B|3B)$")
     instruct: Optional[str] = Field(None, max_length=500)
@@ -398,7 +398,7 @@ class SpeakRequest(BaseModel):
     )
     language: Optional[str] = Field(
         None,
-        pattern="^(zh|en|ja|ko|de|fr|ru|pt|es|it|he|ar|da|el|fi|hi|ms|nl|no|pl|sv|sw|tr)$",
+        pattern="^(zh|en|ja|ko|de|fr|ru|pt|es|it|he|ar|da|el|fi|hi|ms|nl|no|pl|sv|sw|tr|vi)$",
     )
 
 

backend/tests/test_omnivoice_compat.py

@@ -295,6 +295,25 @@ def test_declared_languages_are_resolvable_by_omnivoice():
     assert unresolved == []
 
 
+def test_declared_languages_pass_the_api_pattern():
+    """Regression: the API enum is a second copy of the advertised language list.
+
+    A code can be listed in the engine config and still be rejected with a 422
+    by the ``language`` pattern on the request models, so every advertised code
+    has to be accepted by ``GenerationRequest``.
+    """
+    from pydantic import ValidationError
+
+    from backend.backends import get_model_config
+    from backend.models import GenerationRequest
+
+    for code in get_model_config("omnivoice").languages:
+        GenerationRequest(profile_id="profile", text="Xin chào", language=code)
+
+    with pytest.raises(ValidationError):
+        GenerationRequest(profile_id="profile", text="Xin chào", language="zz")
+
+
 def test_clear_codec_layer_cache_is_safe_when_codec_never_imported():
     from backend.backends.omnivoice_backend import _clear_codec_layer_cache