replace two letter codes with three letter codes

server/gpu/modal/reflector_transcriber.py

@@ -178,7 +178,7 @@ def web():
     @app.post("/transcribe", dependencies=[Depends(apikey_auth)])
     async def transcribe(
         file: UploadFile,
-        source_language: Annotated[str, Body(...)] = "en",
+        source_language: Annotated[str, Body(...)] = "eng",
         timestamp: Annotated[float, Body()] = 0.0
     ) -> TranscriptResponse:
         audio_data = await file.read()

server/gpu/modal/reflector_translator.py

@@ -219,8 +219,8 @@ def web():
     @app.post("/translate", dependencies=[Depends(apikey_auth)])
     async def translate(
             text: str,
-            source_language: Annotated[str, Body(...)] = "en",
+            source_language: Annotated[str, Body(...)] = "eng",
-            target_language: Annotated[str, Body(...)] = "fr",
+            target_language: Annotated[str, Body(...)] = "fra",
     ) -> TranslateResponse:
         func = translatorstub.translate_text.spawn(
             text=text,

server/reflector/processors/audio_transcript_banana.py

@@ -8,20 +8,21 @@ API will be a POST request to TRANSCRIPT_URL:
     "audio_url": "https://...",
     "audio_ext": "wav",
     "timestamp": 123.456
-    "language": "en"
+    "language": "eng"
 }
 ```
 
 """
 
+from pathlib import Path
+
+import httpx
 from reflector.processors.audio_transcript import AudioTranscriptProcessor
 from reflector.processors.audio_transcript_auto import AudioTranscriptAutoProcessor
 from reflector.processors.types import AudioFile, Transcript, Word
 from reflector.settings import settings
 from reflector.storage import Storage
 from reflector.utils.retry import retry
-from pathlib import Path
-import httpx
 
 
 class AudioTranscriptBananaProcessor(AudioTranscriptProcessor):

server/reflector/processors/audio_transcript_modal.py

@@ -5,8 +5,8 @@ API will be a POST request to TRANSCRIPT_URL:
 
 ```form
 "timestamp": 123.456
-"source_language": "en"
+"source_language": "eng"
-"target_language": "en"
+"target_language": "eng"
 "file": <audio file>
 ```
 
@@ -33,7 +33,7 @@ class AudioTranscriptModalProcessor(AudioTranscriptProcessor):
             files = {
                 "file": (data.name, data.fd),
             }
-            source_language = self.get_pref("audio:source_language", "en")
+            source_language = self.get_pref("audio:source_language", "eng")
             json_payload = {"source_language": source_language}
             response = await retry(client.post)(
                 self.transcript_url,

server/reflector/processors/audio_transcript_whisper.py

@@ -1,7 +1,7 @@
+from faster_whisper import WhisperModel
 from reflector.processors.audio_transcript import AudioTranscriptProcessor
 from reflector.processors.audio_transcript_auto import AudioTranscriptAutoProcessor
 from reflector.processors.types import AudioFile, Transcript, Word
-from faster_whisper import WhisperModel
 
 
 class AudioTranscriptWhisperProcessor(AudioTranscriptProcessor):
@@ -14,7 +14,7 @@ class AudioTranscriptWhisperProcessor(AudioTranscriptProcessor):
     async def _transcript(self, data: AudioFile):
         segments, _ = self.model.transcribe(
             data.path.as_posix(),
-            language="en",
+            language="eng",
             beam_size=5,
             # condition_on_previous_text=True,
             word_timestamps=True,

server/reflector/processors/transcript_translator.py

@@ -28,8 +28,8 @@ class TranscriptTranslatorProcessor(Processor):
         # FIXME this should be a processor after, as each user may want
         # different languages
 
-        source_language = self.get_pref("audio:source_language", "en")
+        source_language = self.get_pref("audio:source_language", "eng")
-        target_language = self.get_pref("audio:target_language", "en")
+        target_language = self.get_pref("audio:target_language", "eng")
         if source_language == target_language:
             return
 

server/reflector/processors/types.py

@@ -117,113 +117,64 @@ class FinalTitle(BaseModel):
     title: str
 
 
+# https://github.com/facebookresearch/seamless_communication/tree/main/scripts/m4t/predict#supported-languages
 class TranslationLanguages(BaseModel):
     language_to_id_mapping: dict = {
-        "Afrikaans": "af",
+        "afr": "Afrikaans",
-        "Albanian": "sq",
+        "azj": "North Azerbaijani",
-        "Amharic": "am",
+        "bos": "Bosnian",
-        "Arabic": "ar",
+        "cat": "Catalan",
-        "Armenian": "hy",
+        "ceb": "Cebuano",
-        "Asturian": "ast",
+        "ces": "Czech",
-        "Azerbaijani": "az",
+        "cym": "Welsh",
-        "Bashkir": "ba",
+        "dan": "Danish",
-        "Belarusian": "be",
+        "deu": "German",
-        "Bengali": "bn",
+        "eng": "English",
-        "Bosnian": "bs",
+        "est": "Estonian",
-        "Breton": "br",
+        "eus": "Basque",
-        "Bulgarian": "bg",
+        "fin": "Finnish",
-        "Burmese": "my",
+        "fra": "French",
-        "Catalan; Valencian": "ca",
+        "gaz": "West Central Oromo",
-        "Cebuano": "ceb",
+        "gle": "Irish",
-        "Central Khmer": "km",
+        "glg": "Galician",
-        "Chinese": "zh",
+        "hrv": "Croatian",
-        "Croatian": "hr",
+        "hun": "Hungarian",
-        "Czech": "cs",
+        "ibo": "Igbo",
-        "Danish": "da",
+        "ind": "Indonesian",
-        "Dutch; Flemish": "nl",
+        "isl": "Icelandic",
-        "English": "en",
+        "ita": "Italian",
-        "Estonian": "et",
+        "jav": "Javanese",
-        "Finnish": "fi",
+        "lit": "Lithuanian",
-        "French": "fr",
+        "lug": "Ganda",
-        "Fulah": "ff",
+        "luo": "Luo",
-        "Gaelic; Scottish Gaelic": "gd",
+        "lvs": "Standard Latvian",
-        "Galician": "gl",
+        "mlt": "Maltese",
-        "Ganda": "lg",
+        "nld": "Dutch",
-        "Georgian": "ka",
+        "nno": "Norwegian Nynorsk",
-        "German": "de",
+        "nob": "Norwegian Bokmål",
-        "Greeek": "el",
+        "nya": "Nyanja",
-        "Gujarati": "gu",
+        "pol": "Polish",
-        "Haitian; Haitian Creole": "ht",
+        "por": "Portuguese",
-        "Hausa": "ha",
+        "ron": "Romanian",
-        "Hebrew": "he",
+        "slk": "Slovak",
-        "Hindi": "hi",
+        "slv": "Slovenian",
-        "Hungarian": "hu",
+        "sna": "Shona",
-        "Icelandic": "is",
+        "som": "Somali",
-        "Igbo": "ig",
+        "spa": "Spanish",
-        "Iloko": "ilo",
+        "swe": "Swedish",
-        "Indonesian": "id",
+        "swh": "Swahili",
-        "Irish": "ga",
+        "tgl": "Tagalog",
-        "Italian": "it",
+        "tur": "Turkish",
-        "Japanese": "ja",
+        "uzn": "Northern Uzbek",
-        "Javanese": "jv",
+        "vie": "Vietnamese",
-        "Kannada": "kn",
+        "yor": "Yoruba",
-        "Kazakh": "kk",
+        "zsm": "Standard Malay",
-        "Korean": "ko",
+        "zul": "Zulu",
-        "Lao": "lo",
-        "Latvian": "lv",
-        "Lingala": "ln",
-        "Lithuanian": "lt",
-        "Luxembourgish; Letzeburgesch": "lb",
-        "Macedonian": "mk",
-        "Malagasy": "mg",
-        "Malay": "ms",
-        "Malayalam": "ml",
-        "Marathi": "mr",
-        "Mongolian": "mn",
-        "Nepali": "ne",
-        "Northern Sotho": "ns",
-        "Norwegian": "no",
-        "Occitan": "oc",
-        "Oriya": "or",
-        "Panjabi; Punjabi": "pa",
-        "Persian": "fa",
-        "Polish": "pl",
-        "Portuguese": "pt",
-        "Pushto; Pashto": "ps",
-        "Romanian; Moldavian; Moldovan": "ro",
-        "Russian": "ru",
-        "Serbian": "sr",
-        "Sindhi": "sd",
-        "Sinhala; Sinhalese": "si",
-        "Slovak": "sk",
-        "Slovenian": "sl",
-        "Somali": "so",
-        "Spanish": "es",
-        "Sundanese": "su",
-        "Swahili": "sw",
-        "Swati": "ss",
-        "Swedish": "sv",
-        "Tagalog": "tl",
-        "Tamil": "ta",
-        "Thai": "th",
-        "Tswana": "tn",
-        "Turkish": "tr",
-        "Ukrainian": "uk",
-        "Urdu": "ur",
-        "Uzbek": "uz",
-        "Vietnamese": "vi",
-        "Welsh": "cy",
-        "Western Frisian": "fy",
-        "Wolof": "wo",
-        "Xhosa": "xh",
-        "Yiddish": "yi",
-        "Yoruba": "yo",
-        "Zulu": "zu",
     }
 
     @property
     def supported_languages(self):
-        return self.language_to_id_mapping.values()
+        return self.language_to_id_mapping.keys()
 
     def is_supported(self, lang_id: str) -> bool:
         if lang_id in self.supported_languages:

server/reflector/tools/process.py

@@ -1,7 +1,6 @@
 import asyncio
 
 import av
-
 from reflector.logger import logger
 from reflector.processors import (
     AudioChunkerProcessor,
@@ -23,8 +22,8 @@ async def process_audio_file(
     filename,
     event_callback,
     only_transcript=False,
-    source_language="en",
+    source_language="eng",
-    target_language="en",
+    target_language="eng",
 ):
     # build pipeline for audio processing
     processors = [
@@ -73,8 +72,8 @@ if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     parser.add_argument("source", help="Source file (mp3, wav, mp4...)")
     parser.add_argument("--only-transcript", "-t", action="store_true")
-    parser.add_argument("--source-language", default="en")
+    parser.add_argument("--source-language", default="eng")
-    parser.add_argument("--target-language", default="en")
+    parser.add_argument("--target-language", default="eng")
     parser.add_argument("--output", "-o", help="Output file (output.jsonl)")
     args = parser.parse_args()
 

server/reflector/views/rtc_offer.py

@@ -90,8 +90,8 @@ async def rtc_offer_base(
     event_callback=None,
     event_callback_args=None,
     audio_filename: Path | None = None,
-    source_language: str = "en",
+    source_language: str = "eng",
-    target_language: str = "en",
+    target_language: str = "eng",
 ):
     # build an rtc session
     offer = RTCSessionDescription(sdp=params.sdp, type=params.type)

server/reflector/views/transcripts.py

@@ -87,8 +87,8 @@ class Transcript(BaseModel):
     long_summary: str | None = None
     topics: list[TranscriptTopic] = []
     events: list[TranscriptEvent] = []
-    source_language: str = "en"
+    source_language: str = "eng"
-    target_language: str = "en"
+    target_language: str = "eng"
 
     def add_event(self, event: str, data: BaseModel) -> TranscriptEvent:
         ev = TranscriptEvent(event=event, data=data.model_dump())
@@ -170,8 +170,8 @@ class TranscriptController:
     async def add(
         self,
         name: str,
-        source_language: str = "en",
+        source_language: str = "eng",
-        target_language: str = "en",
+        target_language: str = "eng",
         user_id: str | None = None,
     ):
         transcript = Transcript(
@@ -231,8 +231,8 @@ class GetTranscript(BaseModel):
 
 class CreateTranscript(BaseModel):
     name: str
-    source_language: str = Field("en")
+    source_language: str = Field("eng")
-    target_language: str = Field("en")
+    target_language: str = Field("eng")
 
 
 class UpdateTranscript(BaseModel):

server/tests/conftest.py

@@ -46,7 +46,7 @@ async def dummy_transcript():
 
     class TestAudioTranscriptProcessor(AudioTranscriptProcessor):
         async def _transcript(self, data: AudioFile):
-            source_language = self.get_pref("audio:source_language", "en")
+            source_language = self.get_pref("audio:source_language", "eng")
             print("transcripting", source_language)
             print("pipeline", self.pipeline)
             print("prefs", self.pipeline.prefs)

server/tests/test_transcripts_translation.py

@@ -10,15 +10,15 @@ async def test_transcript_create_default_translation():
         response = await ac.post("/transcripts", json={"name": "test en"})
         assert response.status_code == 200
         assert response.json()["name"] == "test en"
-        assert response.json()["source_language"] == "en"
+        assert response.json()["source_language"] == "eng"
-        assert response.json()["target_language"] == "en"
+        assert response.json()["target_language"] == "eng"
         tid = response.json()["id"]
 
         response = await ac.get(f"/transcripts/{tid}")
         assert response.status_code == 200
         assert response.json()["name"] == "test en"
-        assert response.json()["source_language"] == "en"
+        assert response.json()["source_language"] == "eng"
-        assert response.json()["target_language"] == "en"
+        assert response.json()["target_language"] == "eng"
 
 
 @pytest.mark.asyncio
@@ -31,15 +31,15 @@ async def test_transcript_create_en_fr_translation():
         )
         assert response.status_code == 200
         assert response.json()["name"] == "test en/fr"
-        assert response.json()["source_language"] == "en"
+        assert response.json()["source_language"] == "eng"
-        assert response.json()["target_language"] == "fr"
+        assert response.json()["target_language"] == "fra"
         tid = response.json()["id"]
 
         response = await ac.get(f"/transcripts/{tid}")
         assert response.status_code == 200
         assert response.json()["name"] == "test en/fr"
-        assert response.json()["source_language"] == "en"
+        assert response.json()["source_language"] == "eng"
-        assert response.json()["target_language"] == "fr"
+        assert response.json()["target_language"] == "fra"
 
 
 @pytest.mark.asyncio
@@ -52,12 +52,12 @@ async def test_transcript_create_fr_en_translation():
         )
         assert response.status_code == 200
         assert response.json()["name"] == "test fr/en"
-        assert response.json()["source_language"] == "fr"
+        assert response.json()["source_language"] == "fra"
-        assert response.json()["target_language"] == "en"
+        assert response.json()["target_language"] == "eng"
         tid = response.json()["id"]
 
         response = await ac.get(f"/transcripts/{tid}")
         assert response.status_code == 200
         assert response.json()["name"] == "test fr/en"
-        assert response.json()["source_language"] == "fr"
+        assert response.json()["source_language"] == "fra"
-        assert response.json()["target_language"] == "en"
+        assert response.json()["target_language"] == "eng"