Update app.py

app.py

@@ -1,46 +1,26 @@
-import logging
-from pdfminer.high_level import extract_text
-from huggingface_hub import hf_hub_download
-from balacoon_tts import TTS
 import gradio as gr
-
-# Global TTS module, initialized from a selected model
-tts = None
-
-def read_pdf(file):
-    text = extract_text(file.name)
-    return text
-
-def synthesize_audio(file):
-    text_str = read_pdf(file)
-    if len(text_str) > 1024:
-        text_str = text_str[:1024]
-    samples = tts.synthesize(text_str)
-    return (samples, tts.get_sampling_rate())
+import pdf2speech
+from gtts import gTTS
+from tempfile import NamedTemporaryFile
+import os
+
+def convert_pdf_to_speech(pdf_file):
+    text = pdf2speech.extract_text_from_pdf(pdf_file.name)
+    tts = gTTS(text=text, lang='en')
+    audio_file = NamedTemporaryFile(suffix=".mp3", delete=False)
+    tts.save(audio_file.name)
+    audio_file.close()
+    return audio_file.name
+
+def pdf_to_speech(pdf_file):
+    audio_file_path = convert_pdf_to_speech(pdf_file)
+    return audio_file_path
 
 def main():
-    logging.basicConfig(level=logging.INFO)
-    file_input = gr.inputs.File(label="Upload PDF")
-
-    model_name = "balacoon/tts"  # Set the desired model name here
-    model_path = hf_hub_download(repo_id=model_name, filename="model.pt")
-    global tts
-    tts = TTS(model_path)
-
-    audio = gr.outputs.Audio(label="Generated Audio", type="numpy")
-
-    def generate_audio(file):
-        return synthesize_audio(file)
-
-    iface = gr.Interface(
-        fn=generate_audio,
-        inputs=file_input,
-        outputs=audio,
-        title="PDF TO SPEECH CONVERTER",
-        layout="vertical",
-        debug=True
-    )
+    pdf_input = gr.inputs.File(label="Upload PDF", type="file")
+    audio_output = gr.outputs.Audio(label="Generated Audio")
 
+    iface = gr.Interface(fn=pdf_to_speech, inputs=pdf_input, outputs=audio_output, title="PDF to Speech Converter")
     iface.launch()
 
 if __name__ == "__main__":