selfhostyourtech/reflector
2
0
Fork 0
mirror of https://github.com/Monadical-SAS/reflector.git synced 2025-12-20 20:29:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #288 from Monadical-SAS/feat/lang-dropdown

Browse Source 
Language codes and translation changes
This commit is contained in:
Sara
2023-10-15 15:42:54 +02:00
committed by GitHub
parent 90c6824f52 c1a9005ec3
commit 13075784fe
17 changed files with 621 additions and 281 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
server/gpu/modal/reflector_llm.py
View File

@@ -5,6 +5,7 @@ Reflector GPU backend - LLM
""" """
import json import json
import os import os
import threading
from typing import Optional from typing import Optional
import modal import modal
@@ -67,7 +68,7 @@ llm_image = (
gpu="A100", gpu="A100",
timeout=60 * 5, timeout=60 * 5,
container_idle_timeout=60 * 5, container_idle_timeout=60 * 5,
concurrency_limit=2, allow_concurrent_inputs=15,
image=llm_image, image=llm_image,
) )
class LLM: class LLM:
@@ -108,6 +109,8 @@ class LLM:
self.gen_cfg = gen_cfg self.gen_cfg = gen_cfg
self.GenerationConfig = GenerationConfig self.GenerationConfig = GenerationConfig
self.lock = threading.Lock()
def __exit__(self, *args): def __exit__(self, *args):
print("Exit llm") print("Exit llm")
@@ -123,30 +126,31 @@ class LLM:
gen_cfg = self.gen_cfg gen_cfg = self.gen_cfg
# If a gen_schema is given, conform to gen_schema # If a gen_schema is given, conform to gen_schema
if gen_schema: with self.lock:
import jsonformer if gen_schema:
import jsonformer
print(f"Schema {gen_schema=}") print(f"Schema {gen_schema=}")
jsonformer_llm = jsonformer.Jsonformer( jsonformer_llm = jsonformer.Jsonformer(
model=self.model, model=self.model,
tokenizer=self.tokenizer, tokenizer=self.tokenizer,
json_schema=json.loads(gen_schema), json_schema=json.loads(gen_schema),
prompt=prompt, prompt=prompt,
max_string_token_length=gen_cfg.max_new_tokens max_string_token_length=gen_cfg.max_new_tokens
) )
response = jsonformer_llm() response = jsonformer_llm()
else: else:
# If no gen_schema, perform prompt only generation # If no gen_schema, perform prompt only generation
# tokenize prompt # tokenize prompt
input_ids = self.tokenizer.encode(prompt, return_tensors="pt").to( input_ids = self.tokenizer.encode(prompt, return_tensors="pt").to(
self.model.device self.model.device
) )
output = self.model.generate(input_ids, generation_config=gen_cfg) output = self.model.generate(input_ids, generation_config=gen_cfg)
# decode output # decode output
response = self.tokenizer.decode(output[0].cpu(), skip_special_tokens=True) response = self.tokenizer.decode(output[0].cpu(), skip_special_tokens=True)
response = response[len(prompt):] response = response[len(prompt):]
print(f"Generated {response=}") print(f"Generated {response=}")
return {"text": response} return {"text": response}
@@ -158,6 +162,7 @@ class LLM:
@stub.function( @stub.function(
container_idle_timeout=60 * 10, container_idle_timeout=60 * 10,
timeout=60 * 5, timeout=60 * 5,
allow_concurrent_inputs=45,
secrets=[ secrets=[
Secret.from_name("reflector-gpu"), Secret.from_name("reflector-gpu"),
], ],
@@ -187,7 +192,7 @@ def web():
gen_cfg: Optional[dict] = None gen_cfg: Optional[dict] = None
@app.post("/llm", dependencies=[Depends(apikey_auth)]) @app.post("/llm", dependencies=[Depends(apikey_auth)])
async def llm( def llm(
req: LLMRequest, req: LLMRequest,
): ):
gen_schema = json.dumps(req.gen_schema) if req.gen_schema else None gen_schema = json.dumps(req.gen_schema) if req.gen_schema else None

56
server/gpu/modal/reflector_llm_zephyr.py
View File

@@ -5,6 +5,7 @@ Reflector GPU backend - LLM
""" """
import json import json
import os import os
import threading
from typing import Optional from typing import Optional
import modal import modal
@@ -67,7 +68,7 @@ llm_image = (
gpu="A10G", gpu="A10G",
timeout=60 * 5, timeout=60 * 5,
container_idle_timeout=60 * 5, container_idle_timeout=60 * 5,
concurrency_limit=2, allow_concurrent_inputs=10,
image=llm_image, image=llm_image,
) )
class LLM: class LLM:
@@ -111,6 +112,7 @@ class LLM:
self.tokenizer = tokenizer self.tokenizer = tokenizer
self.gen_cfg = gen_cfg self.gen_cfg = gen_cfg
self.GenerationConfig = GenerationConfig self.GenerationConfig = GenerationConfig
self.lock = threading.Lock()
def __exit__(self, *args): def __exit__(self, *args):
print("Exit llm") print("Exit llm")
@@ -129,33 +131,34 @@ class LLM:
gen_cfg = self.gen_cfg gen_cfg = self.gen_cfg
# If a gen_schema is given, conform to gen_schema # If a gen_schema is given, conform to gen_schema
if gen_schema: with self.lock:
import jsonformer if gen_schema:
import jsonformer
print(f"Schema {gen_schema=}") print(f"Schema {gen_schema=}")
jsonformer_llm = jsonformer.Jsonformer( jsonformer_llm = jsonformer.Jsonformer(
model=self.model, model=self.model,
tokenizer=self.tokenizer, tokenizer=self.tokenizer,
json_schema=json.loads(gen_schema), json_schema=json.loads(gen_schema),
prompt=prompt, prompt=prompt,
max_string_token_length=gen_cfg.max_new_tokens max_string_token_length=gen_cfg.max_new_tokens
) )
response = jsonformer_llm() response = jsonformer_llm()
else: else:
# If no gen_schema, perform prompt only generation # If no gen_schema, perform prompt only generation
# tokenize prompt # tokenize prompt
input_ids = self.tokenizer.encode(prompt, return_tensors="pt").to( input_ids = self.tokenizer.encode(prompt, return_tensors="pt").to(
self.model.device self.model.device
) )
output = self.model.generate(input_ids, generation_config=gen_cfg) output = self.model.generate(input_ids, generation_config=gen_cfg)
# decode output # decode output
response = self.tokenizer.decode(output[0].cpu(), skip_special_tokens=True) response = self.tokenizer.decode(output[0].cpu(), skip_special_tokens=True)
response = response[len(prompt):] response = response[len(prompt):]
response = { response = {
"long_summary": response "long_summary": response
} }
print(f"Generated {response=}") print(f"Generated {response=}")
return {"text": response} return {"text": response}
@@ -167,6 +170,7 @@ class LLM:
@stub.function( @stub.function(
container_idle_timeout=60 * 10, container_idle_timeout=60 * 10,
timeout=60 * 5, timeout=60 * 5,
allow_concurrent_inputs=30,
secrets=[ secrets=[
Secret.from_name("reflector-gpu"), Secret.from_name("reflector-gpu"),
], ],
@@ -196,7 +200,7 @@ def web():
gen_cfg: Optional[dict] = None gen_cfg: Optional[dict] = None
@app.post("/llm", dependencies=[Depends(apikey_auth)]) @app.post("/llm", dependencies=[Depends(apikey_auth)])
async def llm( def llm(
req: LLMRequest, req: LLMRequest,
): ):
gen_schema = json.dumps(req.gen_schema) if req.gen_schema else None gen_schema = json.dumps(req.gen_schema) if req.gen_schema else None

27
server/gpu/modal/reflector_transcriber.py
View File

@@ -5,6 +5,7 @@ Reflector GPU backend - transcriber
import os import os
import tempfile import tempfile
import threading
from modal import Image, Secret, Stub, asgi_app, method from modal import Image, Secret, Stub, asgi_app, method
from pydantic import BaseModel from pydantic import BaseModel
@@ -78,6 +79,7 @@ transcriber_image = (
gpu="A10G", gpu="A10G",
timeout=60 * 5, timeout=60 * 5,
container_idle_timeout=60 * 5, container_idle_timeout=60 * 5,
allow_concurrent_inputs=6,
image=transcriber_image, image=transcriber_image,
) )
class Transcriber: class Transcriber:
@@ -85,6 +87,7 @@ class Transcriber:
import faster_whisper import faster_whisper
import torch import torch
self.lock = threading.Lock()
self.use_gpu = torch.cuda.is_available() self.use_gpu = torch.cuda.is_available()
self.device = "cuda" if self.use_gpu else "cpu" self.device = "cuda" if self.use_gpu else "cpu"
self.model = faster_whisper.WhisperModel( self.model = faster_whisper.WhisperModel(
@@ -106,14 +109,15 @@ class Transcriber:
with tempfile.NamedTemporaryFile("wb+", suffix=f".{audio_suffix}") as fp: with tempfile.NamedTemporaryFile("wb+", suffix=f".{audio_suffix}") as fp:
fp.write(audio_data) fp.write(audio_data)
segments, _ = self.model.transcribe( with self.lock:
fp.name, segments, _ = self.model.transcribe(
language=source_language, fp.name,
beam_size=5, language=source_language,
word_timestamps=True, beam_size=5,
vad_filter=True, word_timestamps=True,
vad_parameters={"min_silence_duration_ms": 500}, vad_filter=True,
) vad_parameters={"min_silence_duration_ms": 500},
)
multilingual_transcript = {} multilingual_transcript = {}
transcript_source_lang = "" transcript_source_lang = ""
@@ -147,6 +151,7 @@ class Transcriber:
@stub.function( @stub.function(
container_idle_timeout=60, container_idle_timeout=60,
timeout=60, timeout=60,
allow_concurrent_inputs=40,
secrets=[ secrets=[
Secret.from_name("reflector-gpu"), Secret.from_name("reflector-gpu"),
], ],
@@ -176,12 +181,12 @@ def web():
result: dict result: dict
@app.post("/transcribe", dependencies=[Depends(apikey_auth)]) @app.post("/transcribe", dependencies=[Depends(apikey_auth)])
async def transcribe( def transcribe(
file: UploadFile, file: UploadFile,
source_language: Annotated[str, Body(...)] = "eng", source_language: Annotated[str, Body(...)] = "en",
timestamp: Annotated[float, Body()] = 0.0 timestamp: Annotated[float, Body()] = 0.0
) -> TranscriptResponse: ) -> TranscriptResponse:
audio_data = await file.read() audio_data = file.file.read()
audio_suffix = file.filename.split(".")[-1] audio_suffix = file.filename.split(".")[-1]
assert audio_suffix in supported_audio_file_types assert audio_suffix in supported_audio_file_types

281
server/gpu/modal/reflector_translator.py
View File

@@ -4,7 +4,7 @@ Reflector GPU backend - transcriber
""" """
import os import os
import tempfile import threading
from modal import Image, Secret, Stub, asgi_app, method from modal import Image, Secret, Stub, asgi_app, method
from pydantic import BaseModel from pydantic import BaseModel
@@ -26,8 +26,11 @@ stub = Stub(name="reflector-translator")
def install_seamless_communication(): def install_seamless_communication():
import os import os
import subprocess import subprocess
initial_dir = os.getcwd() initial_dir = os.getcwd()
subprocess.run(["ssh-keyscan", "-t", "rsa", "github.com", ">>", "~/.ssh/known_hosts"]) subprocess.run(
["ssh-keyscan", "-t", "rsa", "github.com", ">>", "~/.ssh/known_hosts"]
)
subprocess.run(["rm", "-rf", "seamless_communication"]) subprocess.run(["rm", "-rf", "seamless_communication"])
subprocess.run(["git", "clone", SEAMLESS_GITEPO, "." + "/seamless_communication"]) subprocess.run(["git", "clone", SEAMLESS_GITEPO, "." + "/seamless_communication"])
os.chdir("seamless_communication") os.chdir("seamless_communication")
@@ -54,13 +57,13 @@ def configure_seamless_m4t():
ASSETS_DIR: str = "./seamless_communication/src/seamless_communication/assets/cards" ASSETS_DIR: str = "./seamless_communication/src/seamless_communication/assets/cards"
with open(f'{ASSETS_DIR}/seamlessM4T_{SEAMLESSM4T_MODEL_SIZE}.yaml', 'r') as file: with open(f"{ASSETS_DIR}/seamlessM4T_{SEAMLESSM4T_MODEL_SIZE}.yaml", "r") as file:
model_yaml_data = yaml.load(file, Loader=yaml.FullLoader) model_yaml_data = yaml.load(file, Loader=yaml.FullLoader)
with open(f'{ASSETS_DIR}/vocoder_36langs.yaml', 'r') as file: with open(f"{ASSETS_DIR}/vocoder_36langs.yaml", "r") as file:
vocoder_yaml_data = yaml.load(file, Loader=yaml.FullLoader) vocoder_yaml_data = yaml.load(file, Loader=yaml.FullLoader)
with open(f'{ASSETS_DIR}/unity_nllb-100.yaml', 'r') as file: with open(f"{ASSETS_DIR}/unity_nllb-100.yaml", "r") as file:
unity_100_yaml_data = yaml.load(file, Loader=yaml.FullLoader) unity_100_yaml_data = yaml.load(file, Loader=yaml.FullLoader)
with open(f'{ASSETS_DIR}/unity_nllb-200.yaml', 'r') as file: with open(f"{ASSETS_DIR}/unity_nllb-200.yaml", "r") as file:
unity_200_yaml_data = yaml.load(file, Loader=yaml.FullLoader) unity_200_yaml_data = yaml.load(file, Loader=yaml.FullLoader)
model_dir = f"{SEAMLESS_MODEL_DIR}/models--facebook--seamless-m4t-{SEAMLESSM4T_MODEL_SIZE}/snapshots" model_dir = f"{SEAMLESS_MODEL_DIR}/models--facebook--seamless-m4t-{SEAMLESSM4T_MODEL_SIZE}/snapshots"
@@ -69,27 +72,33 @@ def configure_seamless_m4t():
model_name = f"multitask_unity_{SEAMLESSM4T_MODEL_SIZE}.pt" model_name = f"multitask_unity_{SEAMLESSM4T_MODEL_SIZE}.pt"
model_path = os.path.join(os.getcwd(), model_dir, latest_model_version, model_name) model_path = os.path.join(os.getcwd(), model_dir, latest_model_version, model_name)
vocoder_dir = f"{SEAMLESS_MODEL_DIR}/models--facebook--seamless-m4t-vocoder/snapshots" vocoder_dir = (
f"{SEAMLESS_MODEL_DIR}/models--facebook--seamless-m4t-vocoder/snapshots"
)
available_vocoder_versions = os.listdir(vocoder_dir) available_vocoder_versions = os.listdir(vocoder_dir)
latest_vocoder_version = sorted(available_vocoder_versions)[-1] latest_vocoder_version = sorted(available_vocoder_versions)[-1]
vocoder_name = "vocoder_36langs.pt" vocoder_name = "vocoder_36langs.pt"
vocoder_path = os.path.join(os.getcwd(), vocoder_dir, latest_vocoder_version, vocoder_name) vocoder_path = os.path.join(
os.getcwd(), vocoder_dir, latest_vocoder_version, vocoder_name
)
tokenizer_name = "tokenizer.model" tokenizer_name = "tokenizer.model"
tokenizer_path = os.path.join(os.getcwd(), model_dir, latest_model_version, tokenizer_name) tokenizer_path = os.path.join(
os.getcwd(), model_dir, latest_model_version, tokenizer_name
)
model_yaml_data['checkpoint'] = f"file:/{model_path}" model_yaml_data["checkpoint"] = f"file:/{model_path}"
vocoder_yaml_data['checkpoint'] = f"file:/{vocoder_path}" vocoder_yaml_data["checkpoint"] = f"file:/{vocoder_path}"
unity_100_yaml_data['tokenizer'] = f"file:/{tokenizer_path}" unity_100_yaml_data["tokenizer"] = f"file:/{tokenizer_path}"
unity_200_yaml_data['tokenizer'] = f"file:/{tokenizer_path}" unity_200_yaml_data["tokenizer"] = f"file:/{tokenizer_path}"
with open(f'{ASSETS_DIR}/seamlessM4T_{SEAMLESSM4T_MODEL_SIZE}.yaml', 'w') as file: with open(f"{ASSETS_DIR}/seamlessM4T_{SEAMLESSM4T_MODEL_SIZE}.yaml", "w") as file:
yaml.dump(model_yaml_data, file) yaml.dump(model_yaml_data, file)
with open(f'{ASSETS_DIR}/vocoder_36langs.yaml', 'w') as file: with open(f"{ASSETS_DIR}/vocoder_36langs.yaml", "w") as file:
yaml.dump(vocoder_yaml_data, file) yaml.dump(vocoder_yaml_data, file)
with open(f'{ASSETS_DIR}/unity_nllb-100.yaml', 'w') as file: with open(f"{ASSETS_DIR}/unity_nllb-100.yaml", "w") as file:
yaml.dump(unity_100_yaml_data, file) yaml.dump(unity_100_yaml_data, file)
with open(f'{ASSETS_DIR}/unity_nllb-200.yaml', 'w') as file: with open(f"{ASSETS_DIR}/unity_nllb-200.yaml", "w") as file:
yaml.dump(unity_200_yaml_data, file) yaml.dump(unity_200_yaml_data, file)
@@ -109,7 +118,7 @@ transcriber_image = (
"torchaudio", "torchaudio",
"fairseq2", "fairseq2",
"pyyaml", "pyyaml",
"hf-transfer~=0.1" "hf-transfer~=0.1",
) )
.run_function(install_seamless_communication) .run_function(install_seamless_communication)
.run_function(download_seamlessm4t_model) .run_function(download_seamlessm4t_model)
@@ -129,6 +138,7 @@ transcriber_image = (
gpu="A10G", gpu="A10G",
timeout=60 * 5, timeout=60 * 5,
container_idle_timeout=60 * 5, container_idle_timeout=60 * 5,
allow_concurrent_inputs=4,
image=transcriber_image, image=transcriber_image,
) )
class Translator: class Translator:
@@ -136,13 +146,14 @@ class Translator:
import torch import torch
from seamless_communication.models.inference.translator import Translator from seamless_communication.models.inference.translator import Translator
self.lock = threading.Lock()
self.use_gpu = torch.cuda.is_available() self.use_gpu = torch.cuda.is_available()
self.device = "cuda" if self.use_gpu else "cpu" self.device = "cuda" if self.use_gpu else "cpu"
self.translator = Translator( self.translator = Translator(
SEAMLESSM4T_MODEL_CARD_NAME, SEAMLESSM4T_MODEL_CARD_NAME,
SEAMLESSM4T_VOCODER_CARD_NAME, SEAMLESSM4T_VOCODER_CARD_NAME,
torch.device(self.device), torch.device(self.device),
dtype=torch.float32 dtype=torch.float32,
) )
@method() @method()
@@ -156,31 +167,212 @@ class Translator:
""" """
# TODO: Enhance with complete list of lang codes # TODO: Enhance with complete list of lang codes
seamless_lang_code = { seamless_lang_code = {
"en": "eng", # Afrikaans
"fr": "fra" 'af': 'afr',
# Amharic
'am': 'amh',
# Modern Standard Arabic
'ar': 'arb',
# Moroccan Arabic
'ary': 'ary',
# Egyptian Arabic
'arz': 'arz',
# Assamese
'as': 'asm',
# North Azerbaijani
'az': 'azj',
# Belarusian
'be': 'bel',
# Bengali
'bn': 'ben',
# Bosnian
'bs': 'bos',
# Bulgarian
'bg': 'bul',
# Catalan
'ca': 'cat',
# Cebuano
'ceb': 'ceb',
# Czech
'cs': 'ces',
# Central Kurdish
'ku': 'ckb',
# Mandarin Chinese
'cmn': 'cmn_Hant',
# Welsh
'cy': 'cym',
# Danish
'da': 'dan',
# German
'de': 'deu',
# Greek
'el': 'ell',
# English
'en': 'eng',
# Estonian
'et': 'est',
# Basque
'eu': 'eus',
# Finnish
'fi': 'fin',
# French
'fr': 'fra',
# Irish
'ga': 'gle',
# West Central Oromo,
'gaz': 'gaz',
# Galician
'gl': 'glg',
# Gujarati
'gu': 'guj',
# Hebrew
'he': 'heb',
# Hindi
'hi': 'hin',
# Croatian
'hr': 'hrv',
# Hungarian
'hu': 'hun',
# Armenian
'hy': 'hye',
# Igbo
'ig': 'ibo',
# Indonesian
'id': 'ind',
# Icelandic
'is': 'isl',
# Italian
'it': 'ita',
# Javanese
'jv': 'jav',
# Japanese
'ja': 'jpn',
# Kannada
'kn': 'kan',
# Georgian
'ka': 'kat',
# Kazakh
'kk': 'kaz',
# Halh Mongolian
'khk': 'khk',
# Khmer
'km': 'khm',
# Kyrgyz
'ky': 'kir',
# Korean
'ko': 'kor',
# Lao
'lo': 'lao',
# Lithuanian
'lt': 'lit',
# Ganda
'lg': 'lug',
# Luo
'luo': 'luo',
# Standard Latvian
'lv': 'lvs',
# Maithili
'mai': 'mai',
# Malayalam
'ml': 'mal',
# Marathi
'mr': 'mar',
# Macedonian
'mk': 'mkd',
# Maltese
'mt': 'mlt',
# Meitei
'mni': 'mni',
# Burmese
'my': 'mya',
# Dutch
'nl': 'nld',
# Norwegian Nynorsk
'nn': 'nno',
# Norwegian Bokmål
'nb': 'nob',
# Nepali
'ne': 'npi',
# Nyanja
'ny': 'nya',
# Odia
'or': 'ory',
# Punjabi
'pa': 'pan',
# Southern Pashto
'pbt': 'pbt',
# Western Persian
'pes': 'pes',
# Polish
'pl': 'pol',
# Portuguese
'pt': 'por',
# Romanian
'ro': 'ron',
# Russian
'ru': 'rus',
# Slovak
'sk': 'slk',
# Slovenian
'sl': 'slv',
# Shona
'sn': 'sna',
# Sindhi
'sd': 'snd',
# Somali
'so': 'som',
# Spanish
'es': 'spa',
# Serbian
'sr': 'srp',
# Swedish
'sv': 'swe',
# Swahili
'sw': 'swh',
# Tamil
'ta': 'tam',
# Telugu
'te': 'tel',
# Tajik
'tg': 'tgk',
# Tagalog
'tl': 'tgl',
# Thai
'th': 'tha',
# Turkish
'tr': 'tur',
# Ukrainian
'uk': 'ukr',
# Urdu
'ur': 'urd',
# Northern Uzbek
'uz': 'uzn',
# Vietnamese
'vi': 'vie',
# Yoruba
'yo': 'yor',
# Cantonese
'yue': 'yue',
# Standard Malay
'ms': 'zsm',
# Zulu
'zu': 'zul'
} }
return seamless_lang_code.get(lang_code, "eng") return seamless_lang_code.get(lang_code, "eng")
@method() @method()
def translate_text( def translate_text(self, text: str, source_language: str, target_language: str):
self, with self.lock:
text: str, translated_text, _, _ = self.translator.predict(
source_language: str, text,
target_language: str "t2tt",
): src_lang=self.get_seamless_lang_code(source_language),
translated_text, _, _ = self.translator.predict( tgt_lang=self.get_seamless_lang_code(target_language),
text, ngram_filtering=True,
"t2tt", )
src_lang=self.get_seamless_lang_code(source_language), return {"text": {source_language: text, target_language: str(translated_text)}}
tgt_lang=self.get_seamless_lang_code(target_language),
ngram_filtering=True
)
return {
"text": {
source_language: text,
target_language: str(translated_text)
}
}
# ------------------------------------------------------------------- # -------------------------------------------------------------------
# Web API # Web API
# ------------------------------------------------------------------- # -------------------------------------------------------------------
@@ -189,6 +381,7 @@ class Translator:
@stub.function( @stub.function(
container_idle_timeout=60, container_idle_timeout=60,
timeout=60, timeout=60,
allow_concurrent_inputs=40,
secrets=[ secrets=[
Secret.from_name("reflector-gpu"), Secret.from_name("reflector-gpu"),
], ],
@@ -219,8 +412,8 @@ def web():
@app.post("/translate", dependencies=[Depends(apikey_auth)]) @app.post("/translate", dependencies=[Depends(apikey_auth)])
async def translate( async def translate(
text: str, text: str,
source_language: Annotated[str, Body(...)] = "eng", source_language: Annotated[str, Body(...)] = "en",
target_language: Annotated[str, Body(...)] = "fra", target_language: Annotated[str, Body(...)] = "fr",
) -> TranslateResponse: ) -> TranslateResponse:
func = translatorstub.translate_text.spawn( func = translatorstub.translate_text.spawn(
text=text, text=text,
@@ -230,8 +423,4 @@ def web():
result = func.get() result = func.get()
return result return result
@app.post("/warmup", dependencies=[Depends(apikey_auth)])
async def warmup():
return translatorstub.warmup.spawn().get()
return app return app

2
server/reflector/processors/audio_transcript_banana.py
View File

@@ -8,7 +8,7 @@ API will be a POST request to TRANSCRIPT_URL:
"audio_url": "https://...", "audio_url": "https://...",
"audio_ext": "wav", "audio_ext": "wav",
"timestamp": 123.456 "timestamp": 123.456
"language": "eng" "language": "en"
} }
``` ```

6
server/reflector/processors/audio_transcript_modal.py
View File

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

2
server/reflector/processors/audio_transcript_whisper.py
View File

@@ -14,7 +14,7 @@ class AudioTranscriptWhisperProcessor(AudioTranscriptProcessor):
async def _transcript(self, data: AudioFile): async def _transcript(self, data: AudioFile):
segments, _ = self.model.transcribe( segments, _ = self.model.transcribe(
data.path.as_posix(), data.path.as_posix(),
language="eng", language="en",
beam_size=5, beam_size=5,
# condition_on_previous_text=True, # condition_on_previous_text=True,
word_timestamps=True, word_timestamps=True,

2
server/reflector/processors/transcript_final_long_summary.py
View File

@@ -79,7 +79,7 @@ class TranscriptFinalLongSummaryProcessor(Processor):
sentence = str(sentence).strip() sentence = str(sentence).strip()
if sentence.startswith("- "): if sentence.startswith("- "):
sentence.replace("- ", "* ") sentence.replace("- ", "* ")
else: elif not sentence.startswith("*"):
sentence = "* " + sentence sentence = "* " + sentence
sentence += " \n" sentence += " \n"
summary_sentences.append(sentence) summary_sentences.append(sentence)

4
server/reflector/processors/transcript_translator.py
View File

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

244
server/reflector/processors/types.py
View File

@@ -120,56 +120,196 @@ class FinalTitle(BaseModel):
# https://github.com/facebookresearch/seamless_communication/tree/main/scripts/m4t/predict#supported-languages # https://github.com/facebookresearch/seamless_communication/tree/main/scripts/m4t/predict#supported-languages
class TranslationLanguages(BaseModel): class TranslationLanguages(BaseModel):
language_to_id_mapping: dict = { language_to_id_mapping: dict = {
"afr": "Afrikaans", # Afrikaans
"azj": "North Azerbaijani", "af": "afr",
"bos": "Bosnian", # Amharic
"cat": "Catalan", "am": "amh",
"ceb": "Cebuano", # Modern Standard Arabic
"ces": "Czech", "ar": "arb",
"cym": "Welsh", # Moroccan Arabic
"dan": "Danish", "ary": "ary",
"deu": "German", # Egyptian Arabic
"eng": "English", "arz": "arz",
"est": "Estonian", # Assamese
"eus": "Basque", "as": "asm",
"fin": "Finnish", # North Azerbaijani
"fra": "French", "az": "azj",
"gaz": "West Central Oromo", # Belarusian
"gle": "Irish", "be": "bel",
"glg": "Galician", # Bengali
"hrv": "Croatian", "bn": "ben",
"hun": "Hungarian", # Bosnian
"ibo": "Igbo", "bs": "bos",
"ind": "Indonesian", # Bulgarian
"isl": "Icelandic", "bg": "bul",
"ita": "Italian", # Catalan
"jav": "Javanese", "ca": "cat",
"lit": "Lithuanian", # Cebuano
"lug": "Ganda", "ceb": "ceb",
"luo": "Luo", # Czech
"lvs": "Standard Latvian", "cs": "ces",
"mlt": "Maltese", # Central Kurdish
"nld": "Dutch", "ku": "ckb",
"nno": "Norwegian Nynorsk", # Mandarin Chinese
"nob": "Norwegian Bokmål", "cmn": "cmn_Hant",
"nya": "Nyanja", # Welsh
"pol": "Polish", "cy": "cym",
"por": "Portuguese", # Danish
"ron": "Romanian", "da": "dan",
"slk": "Slovak", # German
"slv": "Slovenian", "de": "deu",
"sna": "Shona", # Greek
"som": "Somali", "el": "ell",
"spa": "Spanish", # English
"swe": "Swedish", "en": "eng",
"swh": "Swahili", # Estonian
"tgl": "Tagalog", "et": "est",
"tur": "Turkish", # Basque
"uzn": "Northern Uzbek", "eu": "eus",
"vie": "Vietnamese", # Finnish
"yor": "Yoruba", "fi": "fin",
"zsm": "Standard Malay", # French
"zul": "Zulu", "fr": "fra",
# Irish
"ga": "gle",
# West Central Oromo,
"gaz": "gaz",
# Galician
"gl": "glg",
# Gujarati
"gu": "guj",
# Hebrew
"he": "heb",
# Hindi
"hi": "hin",
# Croatian
"hr": "hrv",
# Hungarian
"hu": "hun",
# Armenian
"hy": "hye",
# Igbo
"ig": "ibo",
# Indonesian
"id": "ind",
# Icelandic
"is": "isl",
# Italian
"it": "ita",
# Javanese
"jv": "jav",
# Japanese
"ja": "jpn",
# Kannada
"kn": "kan",
# Georgian
"ka": "kat",
# Kazakh
"kk": "kaz",
# Halh Mongolian
"khk": "khk",
# Khmer
"km": "khm",
# Kyrgyz
"ky": "kir",
# Korean
"ko": "kor",
# Lao
"lo": "lao",
# Lithuanian
"lt": "lit",
# Ganda
"lg": "lug",
# Luo
"luo": "luo",
# Standard Latvian
"lv": "lvs",
# Maithili
"mai": "mai",
# Malayalam
"ml": "mal",
# Marathi
"mr": "mar",
# Macedonian
"mk": "mkd",
# Maltese
"mt": "mlt",
# Meitei
"mni": "mni",
# Burmese
"my": "mya",
# Dutch
"nl": "nld",
# Norwegian Nynorsk
"nn": "nno",
# Norwegian Bokmål
"nb": "nob",
# Nepali
"ne": "npi",
# Nyanja
"ny": "nya",
# Odia
"or": "ory",
# Punjabi
"pa": "pan",
# Southern Pashto
"pbt": "pbt",
# Western Persian
"pes": "pes",
# Polish
"pl": "pol",
# Portuguese
"pt": "por",
# Romanian
"ro": "ron",
# Russian
"ru": "rus",
# Slovak
"sk": "slk",
# Slovenian
"sl": "slv",
# Shona
"sn": "sna",
# Sindhi
"sd": "snd",
# Somali
"so": "som",
# Spanish
"es": "spa",
# Serbian
"sr": "srp",
# Swedish
"sv": "swe",
# Swahili
"sw": "swh",
# Tamil
"ta": "tam",
# Telugu
"te": "tel",
# Tajik
"tg": "tgk",
# Tagalog
"tl": "tgl",
# Thai
"th": "tha",
# Turkish
"tr": "tur",
# Ukrainian
"uk": "ukr",
# Urdu
"ur": "urd",
# Northern Uzbek
"uz": "uzn",
# Vietnamese
"vi": "vie",
# Yoruba
"yo": "yor",
# Cantonese
"yue": "yue",
# Standard Malay
"ms": "zsm",
# Zulu
"zu": "zul",
} }
@property @property
@@ -177,6 +317,4 @@ class TranslationLanguages(BaseModel):
return self.language_to_id_mapping.keys() return self.language_to_id_mapping.keys()
def is_supported(self, lang_id: str) -> bool: def is_supported(self, lang_id: str) -> bool:
if lang_id in self.supported_languages: return lang_id in self.supported_languages
return True
return False

8
server/reflector/tools/process.py
View File

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

4
server/reflector/views/rtc_offer.py
View File

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

12
server/reflector/views/transcripts.py
View File

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

2
server/tests/conftest.py
View File

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

24
server/tests/test_transcripts_translation.py
View File

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

175
www/app/supportedLanguages.ts
View File

@@ -8,14 +8,18 @@ type LanguageOption = {
}; };
const supportedLanguages: LanguageOption[] = [ const supportedLanguages: LanguageOption[] = [
{ value: "afr", name: "Afrikaans", script: "Latn" },
{ {
value: "amh", value: "af",
name: "Afrikaans",
script: "Latn",
},
{
value: "am",
name: "Amharic", name: "Amharic",
script: "Ethi", script: "Ethi",
}, },
{ {
value: "arb", value: "ar",
name: "Modern Standard Arabic", name: "Modern Standard Arabic",
script: "Arab", script: "Arab",
}, },
@@ -30,37 +34,37 @@ const supportedLanguages: LanguageOption[] = [
script: "Arab", script: "Arab",
}, },
{ {
value: "asm", value: "as",
name: "Assamese", name: "Assamese",
script: "Beng", script: "Beng",
}, },
{ {
value: "azj", value: "az",
name: "North Azerbaijani", name: "North Azerbaijani",
script: "Latn", script: "Latn",
}, },
{ {
value: "bel", value: "be",
name: "Belarusian", name: "Belarusian",
script: "Cyrl", script: "Cyrl",
}, },
{ {
value: "ben", value: "bn",
name: "Bengali", name: "Bengali",
script: "Beng", script: "Beng",
}, },
{ {
value: "bos", value: "bs",
name: "Bosnian", name: "Bosnian",
script: "Latn", script: "Latn",
}, },
{ {
value: "bul", value: "bg",
name: "Bulgarian", name: "Bulgarian",
script: "Cyrl", script: "Cyrl",
}, },
{ {
value: "cat", value: "ca",
name: "Catalan", name: "Catalan",
script: "Latn", script: "Latn",
}, },
@@ -70,12 +74,12 @@ const supportedLanguages: LanguageOption[] = [
script: "Latn", script: "Latn",
}, },
{ {
value: "ces", value: "cs",
name: "Czech", name: "Czech",
script: "Latn", script: "Latn",
}, },
{ {
value: "ckb", value: "ku",
name: "Central Kurdish", name: "Central Kurdish",
script: "Arab", script: "Arab",
}, },
@@ -85,52 +89,47 @@ const supportedLanguages: LanguageOption[] = [
script: "Hans", script: "Hans",
}, },
{ {
value: "cmn_Ha", value: "cy",
name: "Mandarin Chinese",
script: "Hant",
},
{
value: "cym",
name: "Welsh", name: "Welsh",
script: "Latn", script: "Latn",
}, },
{ {
value: "dan", value: "da",
name: "Danish", name: "Danish",
script: "Latn", script: "Latn",
}, },
{ {
value: "deu", value: "de",
name: "German", name: "German",
script: "Latn", script: "Latn",
}, },
{ {
value: "ell", value: "el",
name: "Greek", name: "Greek",
script: "Grek", script: "Grek",
}, },
{ {
value: "eng", value: "en",
name: "English", name: "English",
script: "Latn", script: "Latn",
}, },
{ {
value: "est", value: "et",
name: "Estonian", name: "Estonian",
script: "Latn", script: "Latn",
}, },
{ {
value: "eus", value: "eu",
name: "Basque", name: "Basque",
script: "Latn", script: "Latn",
}, },
{ {
value: "fin", value: "fi",
name: "Finnish", name: "Finnish",
script: "Latn", script: "Latn",
}, },
{ {
value: "fra", value: "fr",
name: "French", name: "French",
script: "Latn", script: "Latn",
}, },
@@ -140,87 +139,87 @@ const supportedLanguages: LanguageOption[] = [
script: "Latn", script: "Latn",
}, },
{ {
value: "gle", value: "ga",
name: "Irish", name: "Irish",
script: "Latn", script: "Latn",
}, },
{ {
value: "glg", value: "gl",
name: "Galician", name: "Galician",
script: "Latn", script: "Latn",
}, },
{ {
value: "guj", value: "gu",
name: "Gujarati", name: "Gujarati",
script: "Gujr", script: "Gujr",
}, },
{ {
value: "heb", value: "he",
name: "Hebrew", name: "Hebrew",
script: "Hebr", script: "Hebr",
}, },
{ {
value: "hin", value: "hi",
name: "Hindi", name: "Hindi",
script: "Deva", script: "Deva",
}, },
{ {
value: "hrv", value: "hr",
name: "Croatian", name: "Croatian",
script: "Latn", script: "Latn",
}, },
{ {
value: "hun", value: "hu",
name: "Hungarian", name: "Hungarian",
script: "Latn", script: "Latn",
}, },
{ {
value: "hye", value: "hy",
name: "Armenian", name: "Armenian",
script: "Armn", script: "Armn",
}, },
{ {
value: "ibo", value: "ig",
name: "Igbo", name: "Igbo",
script: "Latn", script: "Latn",
}, },
{ {
value: "ind", value: "id",
name: "Indonesian", name: "Indonesian",
script: "Latn", script: "Latn",
}, },
{ {
value: "isl", value: "is",
name: "Icelandic", name: "Icelandic",
script: "Latn", script: "Latn",
}, },
{ {
value: "ita", value: "it",
name: "Italian", name: "Italian",
script: "Latn", script: "Latn",
}, },
{ {
value: "jav", value: "jv",
name: "Javanese", name: "Javanese",
script: "Latn", script: "Latn",
}, },
{ {
value: "jpn", value: "ja",
name: "Japanese", name: "Japanese",
script: "Jpan", script: "Jpan",
}, },
{ {
value: "kan", value: "kn",
name: "Kannada", name: "Kannada",
script: "Knda", script: "Knda",
}, },
{ {
value: "kat", value: "ka",
name: "Georgian", name: "Georgian",
script: "Geor", script: "Geor",
}, },
{ {
value: "kaz", value: "kk",
name: "Kazakh", name: "Kazakh",
script: "Cyrl", script: "Cyrl",
}, },
@@ -230,32 +229,32 @@ const supportedLanguages: LanguageOption[] = [
script: "Cyrl", script: "Cyrl",
}, },
{ {
value: "khm", value: "km",
name: "Khmer", name: "Khmer",
script: "Khmr", script: "Khmr",
}, },
{ {
value: "kir", value: "ky",
name: "Kyrgyz", name: "Kyrgyz",
script: "Cyrl", script: "Cyrl",
}, },
{ {
value: "kor", value: "ko",
name: "Korean", name: "Korean",
script: "Kore", script: "Kore",
}, },
{ {
value: "lao", value: "lo",
name: "Lao", name: "Lao",
script: "Laoo", script: "Laoo",
}, },
{ {
value: "lit", value: "lt",
name: "Lithuanian", name: "Lithuanian",
script: "Latn", script: "Latn",
}, },
{ {
value: "lug", value: "lg",
name: "Ganda", name: "Ganda",
script: "Latn", script: "Latn",
}, },
@@ -265,7 +264,7 @@ const supportedLanguages: LanguageOption[] = [
script: "Latn", script: "Latn",
}, },
{ {
value: "lvs", value: "lv",
name: "Standard Latvian", name: "Standard Latvian",
script: "Latn", script: "Latn",
}, },
@@ -275,22 +274,22 @@ const supportedLanguages: LanguageOption[] = [
script: "Deva", script: "Deva",
}, },
{ {
value: "mal", value: "ml",
name: "Malayalam", name: "Malayalam",
script: "Mlym", script: "Mlym",
}, },
{ {
value: "mar", value: "mr",
name: "Marathi", name: "Marathi",
script: "Deva", script: "Deva",
}, },
{ {
value: "mkd", value: "mk",
name: "Macedonian", name: "Macedonian",
script: "Cyrl", script: "Cyrl",
}, },
{ {
value: "mlt", value: "mt",
name: "Maltese", name: "Maltese",
script: "Latn", script: "Latn",
}, },
@@ -300,42 +299,42 @@ const supportedLanguages: LanguageOption[] = [
script: "Beng", script: "Beng",
}, },
{ {
value: "mya", value: "my",
name: "Burmese", name: "Burmese",
script: "Mymr", script: "Mymr",
}, },
{ {
value: "nld", value: "nl",
name: "Dutch", name: "Dutch",
script: "Latn", script: "Latn",
}, },
{ {
value: "nno", value: "nn",
name: "Norwegian Nynorsk", name: "Norwegian Nynorsk",
script: "Latn", script: "Latn",
}, },
{ {
value: "nob", value: "nb",
name: "Norwegian Bokmål", name: "Norwegian Bokmål",
script: "Latn", script: "Latn",
}, },
{ {
value: "npi", value: "ne",
name: "Nepali", name: "Nepali",
script: "Deva", script: "Deva",
}, },
{ {
value: "nya", value: "ny",
name: "Nyanja", name: "Nyanja",
script: "Latn", script: "Latn",
}, },
{ {
value: "ory", value: "or",
name: "Odia", name: "Odia",
script: "Orya", script: "Orya",
}, },
{ {
value: "pan", value: "pa",
name: "Punjabi", name: "Punjabi",
script: "Guru", script: "Guru",
}, },
@@ -350,122 +349,122 @@ const supportedLanguages: LanguageOption[] = [
script: "Arab", script: "Arab",
}, },
{ {
value: "pol", value: "pl",
name: "Polish", name: "Polish",
script: "Latn", script: "Latn",
}, },
{ {
value: "por", value: "pt",
name: "Portuguese", name: "Portuguese",
script: "Latn", script: "Latn",
}, },
{ {
value: "ron", value: "ro",
name: "Romanian", name: "Romanian",
script: "Latn", script: "Latn",
}, },
{ {
value: "rus", value: "ru",
name: "Russian", name: "Russian",
script: "Cyrl", script: "Cyrl",
}, },
{ {
value: "slk", value: "sk",
name: "Slovak", name: "Slovak",
script: "Latn", script: "Latn",
}, },
{ {
value: "slv", value: "sl",
name: "Slovenian", name: "Slovenian",
script: "Latn", script: "Latn",
}, },
{ {
value: "sna", value: "sn",
name: "Shona", name: "Shona",
script: "Latn", script: "Latn",
}, },
{ {
value: "snd", value: "sd",
name: "Sindhi", name: "Sindhi",
script: "Arab", script: "Arab",
}, },
{ {
value: "som", value: "so",
name: "Somali", name: "Somali",
script: "Latn", script: "Latn",
}, },
{ {
value: "spa", value: "es",
name: "Spanish", name: "Spanish",
script: "Latn", script: "Latn",
}, },
{ {
value: "srp", value: "sr",
name: "Serbian", name: "Serbian",
script: "Cyrl", script: "Cyrl",
}, },
{ {
value: "swe", value: "sv",
name: "Swedish", name: "Swedish",
script: "Latn", script: "Latn",
}, },
{ {
value: "swh", value: "sw",
name: "Swahili", name: "Swahili",
script: "Latn", script: "Latn",
}, },
{ {
value: "tam", value: "ta",
name: "Tamil", name: "Tamil",
script: "Taml", script: "Taml",
}, },
{ {
value: "tel", value: "te",
name: "Telugu", name: "Telugu",
script: "Telu", script: "Telu",
}, },
{ {
value: "tgk", value: "tg",
name: "Tajik", name: "Tajik",
script: "Cyrl", script: "Cyrl",
}, },
{ {
value: "tgl", value: "tl",
name: "Tagalog", name: "Tagalog",
script: "Latn", script: "Latn",
}, },
{ {
value: "tha", value: "th",
name: "Thai", name: "Thai",
script: "Thai", script: "Thai",
}, },
{ {
value: "tur", value: "tr",
name: "Turkish", name: "Turkish",
script: "Latn", script: "Latn",
}, },
{ {
value: "ukr", value: "uk",
name: "Ukrainian", name: "Ukrainian",
script: "Cyrl", script: "Cyrl",
}, },
{ {
value: "urd", value: "ur",
name: "Urdu", name: "Urdu",
script: "Arab", script: "Arab",
}, },
{ {
value: "uzn", value: "uz",
name: "Northern Uzbek", name: "Northern Uzbek",
script: "Latn", script: "Latn",
}, },
{ {
value: "vie", value: "vi",
name: "Vietnamese", name: "Vietnamese",
script: "Latn", script: "Latn",
}, },
{ {
value: "yor", value: "yo",
name: "Yoruba", name: "Yoruba",
script: "Latn", script: "Latn",
}, },
@@ -475,12 +474,12 @@ const supportedLanguages: LanguageOption[] = [
script: "Hant", script: "Hant",
}, },
{ {
value: "zsm", value: "ms",
name: "Standard Malay", name: "Standard Malay",
script: "Latn", script: "Latn",
}, },
{ {
value: "zul", value: "zu",
name: "Zulu", name: "Zulu",
script: "Latn", script: "Latn",
}, },

2
www/app/transcripts/createTranscript.ts
View File

@@ -25,7 +25,7 @@ const useCreateTranscript = (): CreateTranscript => {
const requestParameters: V1TranscriptsCreateRequest = { const requestParameters: V1TranscriptsCreateRequest = {
createTranscript: { createTranscript: {
name: params.name || "Weekly All-Hands", // Default name: params.name || "Weekly All-Hands", // Default
targetLanguage: params.targetLanguage || "eng", // Default targetLanguage: params.targetLanguage || "en", // Default
}, },
}; };