Spaces:
Runtime error
Runtime error
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import PyPDF2
|
| 2 |
+
from transformers import pipeline
|
| 3 |
+
import gradio as gr
|
| 4 |
+
|
| 5 |
+
# Function to extract text from PDF
|
| 6 |
+
def extract_text_from_pdf(pdf_file):
|
| 7 |
+
reader = PyPDF2.PdfReader(pdf_file)
|
| 8 |
+
text = ""
|
| 9 |
+
for page in reader.pages:
|
| 10 |
+
if page and page.extract_text():
|
| 11 |
+
text += page.extract_text()
|
| 12 |
+
return text
|
| 13 |
+
|
| 14 |
+
# Load text-to-speech pipeline from Hugging Face
|
| 15 |
+
tts = pipeline("text-to-speech", model="facebook/fastspeech2-en-ljspeech")
|
| 16 |
+
|
| 17 |
+
# Function to convert PDF to audio with no text limit
|
| 18 |
+
def pdf_to_audio(pdf_file):
|
| 19 |
+
text = extract_text_from_pdf(pdf_file)
|
| 20 |
+
if not text.strip():
|
| 21 |
+
return "", "No text found in PDF"
|
| 22 |
+
|
| 23 |
+
audio = tts(text)
|
| 24 |
+
audio_path = "output_audio.wav"
|
| 25 |
+
with open(audio_path, "wb") as f:
|
| 26 |
+
f.write(audio["audio"]) # Hugging Face TTS returns audio data
|
| 27 |
+
|
| 28 |
+
return audio_path, "Audio generated successfully"
|
| 29 |
+
|
| 30 |
+
# Gradio interface
|
| 31 |
+
interface = gr.Interface(
|
| 32 |
+
fn=pdf_to_audio,
|
| 33 |
+
inputs=gr.File(type="file"),
|
| 34 |
+
outputs=[gr.Audio(type="filepath"), gr.Text()]
|
| 35 |
+
)
|
| 36 |
+
|
| 37 |
+
if __name__ == "__main__":
|
| 38 |
+
interface.launch()
|