app.py

import os
import time
import tempfile
from math import floor
from typing import Optional, List, Dict, Any

import torch
import gradio as gr
import yt_dlp as youtube_dl
import numpy as np
from transformers import pipeline
from transformers.pipelines.audio_utils import ffmpeg_read
from punctuators.models import PunctCapSegModelONNX
from stable_whisper import WhisperResult


# configuration
MODEL_NAME = "kotoba-tech/kotoba-whisper-v1.0"
BATCH_SIZE = 16
CHUNK_LENGTH_S = 15
FILE_LIMIT_MB = 1000
YT_LENGTH_LIMIT_S = 3600  # limit to 1 hour YouTube files


# device setting
if torch.cuda.is_available():
    torch_dtype = torch.bfloat16
    device = "cuda:0"
    model_kwargs = {'attn_implementation': 'sdpa'}
else:
    torch_dtype = torch.float32
    device = "cpu"
    model_kwargs = {}

# define the pipeline
pipe = pipeline(
    task="automatic-speech-recognition",
    model=MODEL_NAME,
    chunk_length_s=CHUNK_LENGTH_S,
    batch_size=BATCH_SIZE,
    torch_dtype=torch_dtype,
    device=device,
    model_kwargs=model_kwargs
)


class Punctuator:

    ja_punctuations = ["!", "?", "、", "。"]

    def __init__(self, model: str = "pcs_47lang"):
        self.punctuation_model = PunctCapSegModelONNX.from_pretrained(model)

    def punctuate(self, pipeline_chunk: List[Dict[str, Any]]) -> List[Dict[str, Any]]:

        def validate_punctuation(raw: str, punctuated: str):
            if 'unk' in punctuated:
                return raw
            if punctuated.count("。") > 1:
                ind = punctuated.rfind("。")
                punctuated = punctuated.replace("。", "")
                punctuated = punctuated[:ind] + "。" + punctuated[ind:]
            return punctuated

        text_edit = self.punctuation_model.infer([c['text'] for c in pipeline_chunk])
        return [
            {
                'timestamp': c['timestamp'],
                'text': validate_punctuation(c['text'], "".join(e))
            } for c, e in zip(pipeline_chunk, text_edit)
        ]


PUNCTUATOR = Punctuator()


def _fix_timestamp(sample_rate: int, result: List[Dict[str, Any]], audio: np.ndarray) -> WhisperResult or None:

    def replace_none_ts(parts):
        total_dur = round(audio.shape[-1] / sample_rate, 3)
        _medium_dur = _ts_nonzero_mask = None

        def ts_nonzero_mask() -> np.ndarray:
            nonlocal _ts_nonzero_mask
            if _ts_nonzero_mask is None:
                _ts_nonzero_mask = np.array([(p['end'] or p['start']) is not None for p in parts])
            return _ts_nonzero_mask

        def medium_dur() -> float:
            nonlocal _medium_dur
            if _medium_dur is None:
                nonzero_dus = [p['end'] - p['start'] for p in parts if None not in (p['end'], p['start'])]
                nonzero_durs = np.array(nonzero_dus)
                _medium_dur = np.median(nonzero_durs) * 2 if len(nonzero_durs) else 2.0
            return _medium_dur

        def _curr_max_end(start: float, next_idx: float) -> float:
            max_end = total_dur
            if next_idx != len(parts):
                mask = np.flatnonzero(ts_nonzero_mask()[next_idx:])
                if len(mask):
                    _part = parts[mask[0]+next_idx]
                    max_end = _part['start'] or _part['end']

            new_end = round(start + medium_dur(), 3)
            if new_end > max_end:
                return max_end
            return new_end

        for i, part in enumerate(parts, 1):
            if part['start'] is None:
                is_first = i == 1
                if is_first:
                    new_start = round((part['end'] or 0) - medium_dur(), 3)
                    part['start'] = max(new_start, 0.0)
                else:
                    part['start'] = parts[i - 2]['end']
            if part['end'] is None:
                no_next_start = i == len(parts) or parts[i]['start'] is None
                part['end'] = _curr_max_end(part['start'], i) if no_next_start else parts[i]['start']

    words = [dict(start=word['timestamp'][0], end=word['timestamp'][1], word=word['text']) for word in result]
    replace_none_ts(words)
    return WhisperResult([words], force_order=True, check_sorted=True)


def fix_timestamp(pipeline_output: List[Dict[str, Any]], audio: np.ndarray, sample_rate: int) -> List[Dict[str, Any]]:
    result = _fix_timestamp(sample_rate=sample_rate, audio=audio, result=pipeline_output)
    result.adjust_by_silence(
        audio,
        q_levels=20,
        k_size=5,
        sample_rate=sample_rate,
        min_word_dur=None,
        word_level=True,
        verbose=True,
        nonspeech_error=0.1,
        use_word_position=True
    )
    if result.has_words:
        result.regroup(True)
    return [{"timestamp": [s.start, s.end], "text": s.text} for s in result.segments]


def format_time(start: Optional[float], end: Optional[float]):

    def _format_time(seconds: Optional[float]):
        if seconds is None:
            return "complete    "
        minutes = floor(seconds / 60)
        hours = floor(seconds / 3600)
        seconds = seconds - hours * 3600 - minutes * 60
        m_seconds = floor(round(seconds - floor(seconds), 3) * 10 ** 3)
        seconds = floor(seconds)
        return f'{hours:02}:{minutes:02}:{seconds:02}.{m_seconds:03}'

    return f"[{_format_time(start)}-> {_format_time(end)}]:"


def get_prediction(inputs, prompt: Optional[str], punctuate_text: bool = True, stabilize_timestamp: bool = True):
    generate_kwargs = {"language": "japanese", "task": "transcribe"}
    if prompt:
        generate_kwargs['prompt_ids'] = pipe.tokenizer.get_prompt_ids(prompt, return_tensors='pt').to(device)
    prediction = pipe(inputs, return_timestamps=True, generate_kwargs=generate_kwargs)
    if stabilize_timestamp:
        prediction['chunks'] = fix_timestamp(pipeline_output=prediction['chunks'],
                                             audio=inputs["array"],
                                             sample_rate=inputs["sampling_rate"]
        )
    if punctuate_text:
        prediction['chunks'] = PUNCTUATOR.punctuate(prediction['chunks'])
    text = "".join([c['text'] for c in prediction['chunks']])
    text_timestamped = "\n".join([
        f"{format_time(*c['timestamp'])} {c['text']}" for c in prediction['chunks']
    ])
    return text, text_timestamped


def transcribe(inputs, prompt, punctuate_text, stabilize_timestamp):
    if inputs is None:
        raise gr.Error("No audio file submitted! Please upload or record an audio file before submitting your request.")
    inputs = ffmpeg_read(inputs, pipe.feature_extractor.sampling_rate)
    inputs = {"array": inputs, "sampling_rate": pipe.feature_extractor.sampling_rate}
    return get_prediction(inputs, prompt, punctuate_text, stabilize_timestamp)


def _return_yt_html_embed(yt_url):
    video_id = yt_url.split("?v=")[-1]
    return f'<center> <iframe width="500" height="320" src="https://www.youtube.com/embed/{video_id}"> </iframe> </center>'


def download_yt_audio(yt_url, filename):
    info_loader = youtube_dl.YoutubeDL()
    try:
        info = info_loader.extract_info(yt_url, download=False)
    except youtube_dl.utils.DownloadError as err:
        raise gr.Error(str(err))
    file_length = info["duration_string"]
    file_h_m_s = file_length.split(":")
    file_h_m_s = [int(sub_length) for sub_length in file_h_m_s]
    if len(file_h_m_s) == 1:
        file_h_m_s.insert(0, 0)
    if len(file_h_m_s) == 2:
        file_h_m_s.insert(0, 0)
    file_length_s = file_h_m_s[0] * 3600 + file_h_m_s[1] * 60 + file_h_m_s[2]
    if file_length_s > YT_LENGTH_LIMIT_S:
        yt_length_limit_hms = time.strftime("%HH:%MM:%SS", time.gmtime(YT_LENGTH_LIMIT_S))
        file_length_hms = time.strftime("%HH:%MM:%SS", time.gmtime(file_length_s))
        raise gr.Error(f"Maximum YouTube length is {yt_length_limit_hms}, got {file_length_hms} YouTube video.")
    ydl_opts = {"outtmpl": filename, "format": "worstvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best"}
    with youtube_dl.YoutubeDL(ydl_opts) as ydl:
        try:
            ydl.download([yt_url])
        except youtube_dl.utils.ExtractorError as err:
            raise gr.Error(str(err))


def yt_transcribe(yt_url, prompt, punctuate_text: bool = True, stabilize_timestamp: bool = True):
    html_embed_str = _return_yt_html_embed(yt_url)
    with tempfile.TemporaryDirectory() as tmpdirname:
        filepath = os.path.join(tmpdirname, "video.mp4")
        download_yt_audio(yt_url, filepath)
        with open(filepath, "rb") as f:
            inputs = f.read()
    inputs = ffmpeg_read(inputs, pipe.feature_extractor.sampling_rate)
    inputs = {"array": inputs, "sampling_rate": pipe.feature_extractor.sampling_rate}
    text, text_timestamped = get_prediction(inputs, prompt, punctuate_text, stabilize_timestamp)
    return html_embed_str, text, text_timestamped


demo = gr.Blocks()
mf_transcribe = gr.Interface(
    fn=transcribe,
    inputs=[
        gr.inputs.Audio(source="microphone", type="filepath", optional=True),
        gr.inputs.Textbox(lines=1, placeholder="Prompt", optional=True),
        gr.inputs.Checkbox(default=True, label="Add punctuations"),
        gr.inputs.Checkbox(default=True, label="Stabilize timestamp")
    ],
    outputs=["text", "text"],
    layout="horizontal",
    theme="huggingface",
    title=f"Transcribe Audio with {os.path.basename(MODEL_NAME)}",
    description=f"Transcribe long-form microphone or audio inputs with the click of a button! Demo uses the Kotoba-Whisper checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files of arbitrary length.",
    allow_flagging="never",
)

file_transcribe = gr.Interface(
    fn=transcribe,
    inputs=[
        gr.inputs.Audio(source="upload", type="filepath", optional=True, label="Audio file"),
        gr.inputs.Textbox(lines=1, placeholder="Prompt", optional=True),
        gr.inputs.Checkbox(default=True, label="Add punctuations"),
        gr.inputs.Checkbox(default=True, label="Stabilize timestamp")
    ],
    outputs=["text", "text"],
    layout="horizontal",
    theme="huggingface",
    title=f"Transcribe Audio with {os.path.basename(MODEL_NAME)}",
    description=f"Transcribe long-form microphone or audio inputs with the click of a button! Demo uses Kotoba-Whisper checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files of arbitrary length.",
    allow_flagging="never",
)
yt_transcribe = gr.Interface(
    fn=yt_transcribe,
    inputs=[
        gr.inputs.Textbox(lines=1, placeholder="Paste the URL to a YouTube video here", label="YouTube URL"),
        gr.inputs.Textbox(lines=1, placeholder="Prompt", optional=True),
        gr.inputs.Checkbox(default=True, label="Add punctuations"),
        gr.inputs.Checkbox(default=True, label="Stabilize timestamp")
    ],
    outputs=["html", "text", "text"],
    layout="horizontal",
    theme="huggingface",
    title=f"Transcribe YouTube with {os.path.basename(MODEL_NAME)}",
    description=f"Transcribe long-form YouTube videos with the click of a button! Demo uses Kotoba-Whisper checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe video files of arbitrary length.",
    allow_flagging="never",
)

with demo:
    gr.TabbedInterface([mf_transcribe, file_transcribe, yt_transcribe], ["Microphone", "Audio file", "YouTube"])

demo.launch(enable_queue=True)