Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -13,11 +13,16 @@ openai.api_key = os.getenv('OPENAI_API_KEY')
|
|
| 13 |
# Function to transcribe audio using OpenAI Whisper
|
| 14 |
def transcribe_audio(audio_file):
|
| 15 |
try:
|
| 16 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
buffer = io.BytesIO()
|
| 18 |
audio.export(buffer, format="wav")
|
| 19 |
buffer.seek(0)
|
| 20 |
buffer.name = "audio.wav"
|
|
|
|
| 21 |
response = openai.Audio.transcribe(
|
| 22 |
"whisper-1",
|
| 23 |
file=buffer,
|
|
@@ -30,15 +35,23 @@ def transcribe_audio(audio_file):
|
|
| 30 |
|
| 31 |
# Function to extract text from PDF
|
| 32 |
def extract_text_from_pdf(pdf_file):
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
|
| 39 |
# Function to get YouTube transcript
|
| 40 |
def get_youtube_transcript(url):
|
| 41 |
try:
|
|
|
|
| 42 |
if "watch?v=" in url:
|
| 43 |
video_id = url.split("watch?v=")[1].split("&")[0]
|
| 44 |
elif "youtu.be/" in url:
|
|
@@ -46,9 +59,11 @@ def get_youtube_transcript(url):
|
|
| 46 |
else:
|
| 47 |
st.error("Invalid YouTube URL.")
|
| 48 |
return None
|
|
|
|
| 49 |
transcript_list = YouTubeTranscriptApi.list_transcripts(video_id)
|
| 50 |
transcript = transcript_list.find_transcript(['en'])
|
| 51 |
transcript_data = transcript.fetch()
|
|
|
|
| 52 |
transcription_text = " ".join([entry['text'] for entry in transcript_data])
|
| 53 |
return transcription_text
|
| 54 |
except Exception as e:
|
|
@@ -77,7 +92,8 @@ def create_pdf(notes):
|
|
| 77 |
|
| 78 |
# Main app
|
| 79 |
def main():
|
| 80 |
-
st.
|
|
|
|
| 81 |
|
| 82 |
st.markdown("---")
|
| 83 |
st.subheader("Upload your file:")
|
|
@@ -101,6 +117,9 @@ def main():
|
|
| 101 |
return
|
| 102 |
elif input_type == "PDF Document" and pdf_input:
|
| 103 |
transcription_text = extract_text_from_pdf(pdf_input)
|
|
|
|
|
|
|
|
|
|
| 104 |
elif input_type == "YouTube URL" and youtube_input:
|
| 105 |
transcription_text = get_youtube_transcript(youtube_input)
|
| 106 |
if not transcription_text:
|
|
|
|
| 13 |
# Function to transcribe audio using OpenAI Whisper
|
| 14 |
def transcribe_audio(audio_file):
|
| 15 |
try:
|
| 16 |
+
# Read the bytes from the uploaded audio file
|
| 17 |
+
audio_bytes = audio_file.read()
|
| 18 |
+
# Use io.BytesIO to create a file-like object
|
| 19 |
+
audio = AudioSegment.from_file(io.BytesIO(audio_bytes))
|
| 20 |
+
# Convert to WAV and prepare for transcription
|
| 21 |
buffer = io.BytesIO()
|
| 22 |
audio.export(buffer, format="wav")
|
| 23 |
buffer.seek(0)
|
| 24 |
buffer.name = "audio.wav"
|
| 25 |
+
# Transcribe audio using OpenAI Whisper
|
| 26 |
response = openai.Audio.transcribe(
|
| 27 |
"whisper-1",
|
| 28 |
file=buffer,
|
|
|
|
| 35 |
|
| 36 |
# Function to extract text from PDF
|
| 37 |
def extract_text_from_pdf(pdf_file):
|
| 38 |
+
try:
|
| 39 |
+
# Read the bytes from the uploaded PDF file
|
| 40 |
+
pdf_bytes = pdf_file.read()
|
| 41 |
+
# Use io.BytesIO to create a file-like object
|
| 42 |
+
reader = PyPDF2.PdfReader(io.BytesIO(pdf_bytes))
|
| 43 |
+
text = ""
|
| 44 |
+
for page in reader.pages:
|
| 45 |
+
text += page.extract_text() + "\n"
|
| 46 |
+
return text
|
| 47 |
+
except Exception as e:
|
| 48 |
+
st.error(f"Error processing PDF: {str(e)}")
|
| 49 |
+
return None
|
| 50 |
|
| 51 |
# Function to get YouTube transcript
|
| 52 |
def get_youtube_transcript(url):
|
| 53 |
try:
|
| 54 |
+
# Extract video ID from URL
|
| 55 |
if "watch?v=" in url:
|
| 56 |
video_id = url.split("watch?v=")[1].split("&")[0]
|
| 57 |
elif "youtu.be/" in url:
|
|
|
|
| 59 |
else:
|
| 60 |
st.error("Invalid YouTube URL.")
|
| 61 |
return None
|
| 62 |
+
# Fetch transcript
|
| 63 |
transcript_list = YouTubeTranscriptApi.list_transcripts(video_id)
|
| 64 |
transcript = transcript_list.find_transcript(['en'])
|
| 65 |
transcript_data = transcript.fetch()
|
| 66 |
+
# Combine transcript texts
|
| 67 |
transcription_text = " ".join([entry['text'] for entry in transcript_data])
|
| 68 |
return transcription_text
|
| 69 |
except Exception as e:
|
|
|
|
| 92 |
|
| 93 |
# Main app
|
| 94 |
def main():
|
| 95 |
+
st.set_page_config(layout="wide")
|
| 96 |
+
st.markdown("<h1 style='text-align: center;'>AI Notes Generation Bot 🤖</h1>", unsafe_allow_html=True)
|
| 97 |
|
| 98 |
st.markdown("---")
|
| 99 |
st.subheader("Upload your file:")
|
|
|
|
| 117 |
return
|
| 118 |
elif input_type == "PDF Document" and pdf_input:
|
| 119 |
transcription_text = extract_text_from_pdf(pdf_input)
|
| 120 |
+
if not transcription_text:
|
| 121 |
+
st.error("Failed to extract text from PDF.")
|
| 122 |
+
return
|
| 123 |
elif input_type == "YouTube URL" and youtube_input:
|
| 124 |
transcription_text = get_youtube_transcript(youtube_input)
|
| 125 |
if not transcription_text:
|