Spaces:
Paused
Paused
Commit
·
29685bd
1
Parent(s):
d0775b4
adding the transcriber downloader code
Browse files- app.py +52 -0
- requirements.txt +3 -0
app.py
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import yt_dlp as ydlp
|
| 3 |
+
from whispercpp import Whisper
|
| 4 |
+
|
| 5 |
+
def download_audio(youtube_url, output_folder='.'):
|
| 6 |
+
ydl_opts = {
|
| 7 |
+
'format': 'bestaudio/best',
|
| 8 |
+
'postprocessors': [{
|
| 9 |
+
'key': 'FFmpegExtractAudio',
|
| 10 |
+
'preferredcodec': 'wav',
|
| 11 |
+
'preferredquality': '192',
|
| 12 |
+
}],
|
| 13 |
+
'outtmpl': f'{output_folder}/audio',
|
| 14 |
+
}
|
| 15 |
+
|
| 16 |
+
with ydlp.YoutubeDL(ydl_opts) as ydl:
|
| 17 |
+
ydl.download([youtube_url])
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
w = Whisper('tiny')
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
def process_general_transcription(transcription):
|
| 24 |
+
formatted_transcription = []
|
| 25 |
+
|
| 26 |
+
for line in transcription:
|
| 27 |
+
if line.startswith('[') and line.endswith(']'):
|
| 28 |
+
formatted_transcription.append(f'\n--- {line[1:-1].upper()} ---\n')
|
| 29 |
+
else:
|
| 30 |
+
formatted_transcription.append(line)
|
| 31 |
+
|
| 32 |
+
transcript_str = "\n".join(formatted_transcription)
|
| 33 |
+
|
| 34 |
+
return transcript_str
|
| 35 |
+
def transcribe_youtube(youtube_url):
|
| 36 |
+
download_audio(youtube_url)
|
| 37 |
+
result = w.transcribe("audio.wav")
|
| 38 |
+
text = w.extract_text(result)
|
| 39 |
+
return process_general_transcription(text)
|
| 40 |
+
with gr.Blocks() as demo:
|
| 41 |
+
gr.Markdown(
|
| 42 |
+
"""
|
| 43 |
+
# CPP Whisperer - Transcribe YouTube Videos
|
| 44 |
+
|
| 45 |
+
""")
|
| 46 |
+
inp = gr.Textbox(placeholder="Insert YT Url here")
|
| 47 |
+
result_button_transcribe = gr.Button('Transcribe')
|
| 48 |
+
out = gr.Textbox()
|
| 49 |
+
result_button_transcribe.click(transcribe_youtube, inputs = inp, outputs = out)
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
demo.launch()
|
requirements.txt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
git+https://github.com/stlukey/whispercpp.py
|
| 2 |
+
gradio
|
| 3 |
+
yt_dlp
|