videopython 0.26.0__tar.gz → 0.26.1__tar.gz

{videopython-0.26.0 → videopython-0.26.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: videopython
-Version: 0.26.0
+Version: 0.26.1
 Summary: Minimal video generation and processing library.
 Project-URL: Homepage, https://videopython.com
 Project-URL: Repository, https://github.com/bartwojtowicz/videopython/

{videopython-0.26.0 → videopython-0.26.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "videopython"
-version = "0.26.0"
+version = "0.26.1"
 description = "Minimal video generation and processing library."
 authors = [
     { name = "Bartosz Wójtowicz", email = "bartoszwojtowicz@outlook.com" },

{videopython-0.26.0 → videopython-0.26.1}/src/videopython/ai/dubbing/dubber.py

@@ -36,12 +36,15 @@ class VideoDubber:
         voice_clone: bool = True,
         enable_diarization: bool = False,
         progress_callback: Callable[[str, float], None] | None = None,
+        transcription: Any = None,
     ) -> DubbingResult:
         """Dub a video into a target language.
 
         Args:
             enable_diarization: Enable speaker diarization to clone each speaker's
                 voice separately. Requires additional VRAM for the diarization model.
+            transcription: Optional pre-computed Transcription object. When provided,
+                the internal Whisper transcription step is skipped.
         """
         if self._local_pipeline is None:
             self._init_local_pipeline()
@@ -54,6 +57,7 @@ class VideoDubber:
             voice_clone=voice_clone,
             enable_diarization=enable_diarization,
             progress_callback=progress_callback,
+            transcription=transcription,
         )
 
     def dub_and_replace(
@@ -65,8 +69,14 @@ class VideoDubber:
         voice_clone: bool = True,
         enable_diarization: bool = False,
         progress_callback: Callable[[str, float], None] | None = None,
+        transcription: Any = None,
     ) -> Video:
-        """Dub a video and return a new video with the dubbed audio."""
+        """Dub a video and return a new video with the dubbed audio.
+
+        Args:
+            transcription: Optional pre-computed Transcription object. When provided,
+                the internal Whisper transcription step is skipped.
+        """
         result = self.dub(
             video=video,
             target_lang=target_lang,
@@ -75,6 +85,7 @@ class VideoDubber:
             voice_clone=voice_clone,
             enable_diarization=enable_diarization,
             progress_callback=progress_callback,
+            transcription=transcription,
         )
         return video.add_audio(result.dubbed_audio, overlay=False)
 

{videopython-0.26.0 → videopython-0.26.1}/src/videopython/ai/dubbing/pipeline.py

@@ -114,21 +114,33 @@ class LocalDubbingPipeline:
         voice_clone: bool = True,
         enable_diarization: bool = False,
         progress_callback: Callable[[str, float], None] | None = None,
+        transcription: Any | None = None,
     ) -> DubbingResult:
-        """Process a video through the local dubbing pipeline."""
+        """Process a video through the local dubbing pipeline.
+
+        Args:
+            transcription: Optional pre-computed Transcription object. When provided,
+                the internal Whisper transcription step is skipped (saving time and VRAM).
+                Must be a ``videopython.base.text.transcription.Transcription`` instance
+                with populated ``segments``.
+        """
         from videopython.base.audio import Audio
 
         def report_progress(stage: str, progress: float) -> None:
             if progress_callback:
                 progress_callback(stage, progress)
 
-        report_progress("Transcribing audio", 0.05)
-        if self._transcriber is None or self._transcriber_diarization != enable_diarization:
-            self._init_transcriber(enable_diarization=enable_diarization)
-            self._transcriber_diarization = enable_diarization
-
         source_audio = video.audio
-        transcription = self._transcriber.transcribe(source_audio)
+
+        if transcription is not None:
+            report_progress("Using provided transcription", 0.05)
+        else:
+            report_progress("Transcribing audio", 0.05)
+            if self._transcriber is None or self._transcriber_diarization != enable_diarization:
+                self._init_transcriber(enable_diarization=enable_diarization)
+                self._transcriber_diarization = enable_diarization
+
+            transcription = self._transcriber.transcribe(source_audio)
 
         if not transcription.segments:
             return DubbingResult(

{videopython-0.26.0 → videopython-0.26.1}/src/videopython/ai/generation/translation.py

@@ -48,6 +48,15 @@ LANGUAGE_NAMES = {
 class TextTranslator:
     """Translates text between languages using local seq2seq models."""
 
+    # Languages without a direct opus-mt-{src}-{tgt} model. Maps (source, target)
+    # to an alternative HuggingFace model identifier.
+    _MODEL_OVERRIDES: dict[tuple[str, str], str] = {
+        ("en", "pt"): "Helsinki-NLP/opus-mt-tc-big-en-pt",
+        ("en", "ko"): "Helsinki-NLP/opus-mt-tc-big-en-ko",
+        ("en", "ja"): "Helsinki-NLP/opus-mt-en-jap",
+        ("en", "pl"): "Helsinki-NLP/opus-mt-en-zlw",
+    }
+
     def __init__(self, model_name: str | None = None, device: str | None = None):
         self.model_name = model_name
         self.device = device
@@ -58,6 +67,9 @@ class TextTranslator:
     def _get_local_model_name(self, source_lang: str, target_lang: str) -> str:
         if self.model_name:
             return self.model_name
+        override = self._MODEL_OVERRIDES.get((source_lang, target_lang))
+        if override:
+            return override
         return f"Helsinki-NLP/opus-mt-{source_lang}-{target_lang}"
 
     def _init_local(self, source_lang: str, target_lang: str) -> None: