whisper_ft_ITA / app.py
SaladSlayer00's picture
the new app with translation support, youtube and summarizaiton, pdf, tts
a385d52
raw history blame
No virus
4.28 kB
import gradio as gr
import pytube as pt
from transformers import pipeline
import os
from huggingface_hub import HfFolder
from gtts import gTTS
from fpdf import FPDF
from pdfminer.high_level import extract_text
# Initialize pipelines for transcription, summarization, and translation
transcription_pipe = pipeline(model="SaladSlayer00/another_local", token=HfFolder.get_token())
summarizer = pipeline("summarization", model="it5/it5-efficient-small-el32-news-summarization")
translator = pipeline("translation", model="Helsinki-NLP/opus-mt-it-en")
def process_audio(file_path):
text = transcription_pipe(file_path)["text"]
summary = summarizer(text, min_length=25, max_length=50)[0]["summary_text"]
translation = translator(summary)[0]["translation_text"]
return text, summary, translation
def download_youtube_audio(yt_url):
yt = pt.YouTube(yt_url)
stream = yt.streams.filter(only_audio=True).first()
file_path = stream.download(filename="temp_audio.mp3")
return file_path
def youtube_transcription(yt_url):
audio_path = download_youtube_audio(yt_url)
results = process_audio(audio_path)
os.remove(audio_path) # Clean up the downloaded file
return results
def transcribe_and_process(rec=None, file=None):
if rec is not None:
audio = rec
elif file is not None:
audio = file
else:
return "Provide a recording or a file."
return process_audio(audio)
def save_text_to_pdf(text, filename="output.pdf"):
# Create instance of FPDF class
pdf = FPDF()
# Add a page
pdf.add_page()
# Set font: Arial, bold, 12
pdf.set_font("Arial", size=12)
# Add a cell
pdf.multi_cell(0, 10, text)
# Save the pdf with name .pdf
pdf.output(filename)
return filename
def pdf_to_text(file_path):
text = extract_text(file_path)
audio_file = "tts_audio.wav"
myobj = gTTS(text=text, lang='en', slow=False)
myobj.save(audio_file)
return audio_file
def audio_to_pdf(file_path):
text, summary, translation = process_audio(file_path)
pdf_file = save_text_to_pdf(translation)
tts_audio_file = pdf_to_text(pdf_file) # Generate TTS audio from the PDF
return translation, pdf_file, tts_audio_file
def pdf_to_audio(file_path):
text = extract_text(file_path)
myobj = gTTS(text=text, lang='en', slow=False)
audio_file = "output_audio.wav"
myobj.save(audio_file)
return audio_file
app = gr.Blocks()
with app:
gr.Markdown("### Whisper Small Italian Transcription, Summarization, and Translation")
gr.Markdown("Talk, upload an audio file or enter a YouTube URL for processing.")
with gr.Tab("Audio Processing"):
with gr.Row():
audio_input = gr.Audio(label="Upload Audio or Record", type="filepath")
audio_process_button = gr.Button("Process Audio")
audio_transcription, audio_summary, audio_translation = gr.Textbox(label="Transcription"), gr.Textbox(label="Summary"), gr.Textbox(label="Translation")
audio_process_button.click(fn=transcribe_and_process, inputs=audio_input, outputs=[audio_transcription, audio_summary, audio_translation])
with gr.Tab("YouTube Processing"):
with gr.Row():
yt_input = gr.Textbox(label="YouTube URL")
yt_process_button = gr.Button("Process YouTube Video")
yt_transcription, yt_summary, yt_translation = gr.Textbox(label="Transcription"), gr.Textbox(label="Summary"), gr.Textbox(label="Translation")
yt_process_button.click(fn=youtube_transcription, inputs=yt_input, outputs=[yt_transcription, yt_summary, yt_translation])
with gr.Tab("Italian Audio to English PDF"):
with gr.Row():
audio_input = gr.Audio(label="Upload Italian Audio", type="filepath")
translate_process_button = gr.Button("Translate and Save as PDF")
translation_textbox, pdf_download, tts_audio = gr.Textbox(label="Translation"), gr.File(label="Download PDF"), gr.Audio(label="TTS Audio")
translate_process_button.click(fn=audio_to_pdf, inputs=audio_input, outputs=[translation_textbox, pdf_download, tts_audio])
(app.launch())