Spaces:
Sleeping
Sleeping
app.py
DELETED
@@ -1,90 +0,0 @@
|
|
1 |
-
import streamlit as st
|
2 |
-
from pytube import YouTube
|
3 |
-
from moviepy.editor import AudioFileClip
|
4 |
-
import speech_recognition as sr
|
5 |
-
from transformers import pipeline
|
6 |
-
from pydub import AudioSegment, silence
|
7 |
-
import os
|
8 |
-
|
9 |
-
def download_and_extract_audio(youtube_link):
|
10 |
-
try:
|
11 |
-
yt = YouTube(youtube_link)
|
12 |
-
video = yt.streams.filter(only_audio=True).first()
|
13 |
-
out_file = video.download(output_path=".")
|
14 |
-
base, ext = os.path.splitext(out_file)
|
15 |
-
new_file = base + '.mp3'
|
16 |
-
os.rename(out_file, new_file)
|
17 |
-
|
18 |
-
# Convert MP3 to WAV using moviepy
|
19 |
-
audio_clip = AudioFileClip(new_file)
|
20 |
-
wav_file = base + '.wav'
|
21 |
-
audio_clip.write_audiofile(wav_file, codec='pcm_s16le') # codec for WAV format
|
22 |
-
audio_clip.close()
|
23 |
-
return wav_file
|
24 |
-
except Exception as e:
|
25 |
-
st.error(f"Error downloading and extracting audio: {e}")
|
26 |
-
return None
|
27 |
-
|
28 |
-
def transcribe_audio(audio_path):
|
29 |
-
try:
|
30 |
-
recognizer = sr.Recognizer()
|
31 |
-
audio = AudioSegment.from_wav(audio_path)
|
32 |
-
|
33 |
-
# Split the audio into non-silent chunks
|
34 |
-
audio_chunks = silence.split_on_silence(audio, min_silence_len=1000, silence_thresh=-40)
|
35 |
-
|
36 |
-
full_text = ""
|
37 |
-
for i, chunk in enumerate(audio_chunks):
|
38 |
-
chunk.export(f"/tmp/chunk{i}.wav", format="wav")
|
39 |
-
with sr.AudioFile(f"/tmp/chunk{i}.wav") as source:
|
40 |
-
audio_data = recognizer.record(source)
|
41 |
-
try:
|
42 |
-
text = recognizer.recognize_google(audio_data)
|
43 |
-
full_text += " " + text
|
44 |
-
except sr.UnknownValueError:
|
45 |
-
# Skip segments that the recognizer can't understand
|
46 |
-
continue
|
47 |
-
|
48 |
-
return full_text.strip()
|
49 |
-
except Exception as e:
|
50 |
-
st.error(f"Error transcribing audio: {e}")
|
51 |
-
return None
|
52 |
-
|
53 |
-
def summarize_text(text):
|
54 |
-
try:
|
55 |
-
summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
|
56 |
-
summary = summarizer(text, max_length=130, min_length=30, do_sample=False)
|
57 |
-
return summary[0]['summary_text']
|
58 |
-
except Exception as e:
|
59 |
-
st.error(f"Error summarizing text: {e}")
|
60 |
-
return None
|
61 |
-
|
62 |
-
def main():
|
63 |
-
st.title("YouTube Video Summary")
|
64 |
-
youtube_link = st.text_input("Enter YouTube Video Link:")
|
65 |
-
if st.button("Summarize"):
|
66 |
-
if youtube_link:
|
67 |
-
with st.spinner('Downloading and extracting audio...'):
|
68 |
-
audio_path = download_and_extract_audio(youtube_link)
|
69 |
-
if audio_path is None:
|
70 |
-
return
|
71 |
-
|
72 |
-
with st.spinner('Transcribing audio to text...'):
|
73 |
-
transcription = transcribe_audio(audio_path)
|
74 |
-
if transcription is None:
|
75 |
-
return
|
76 |
-
|
77 |
-
with st.spinner('Summarizing transcription...'):
|
78 |
-
summary = summarize_text(transcription)
|
79 |
-
if summary is None:
|
80 |
-
return
|
81 |
-
|
82 |
-
st.subheader("Transcription:")
|
83 |
-
st.write(transcription)
|
84 |
-
st.subheader("Summary:")
|
85 |
-
st.write(summary)
|
86 |
-
else:
|
87 |
-
st.error("Please enter a valid YouTube video link.")
|
88 |
-
|
89 |
-
if __name__ == '__main__':
|
90 |
-
main()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|