Rename pdf_to_audio.py to app.py

pdf_to_audio.py → app.py

@@ -75,4 +75,31 @@ if __name__ == "__main__":
     tts = QuantizedSpeechT5TTSPipe()
     sample_text = 'Hello world! This is a test.'
     result = tts.generate(sample_text)
-    print(f'Generated {len(result)} audio files from "{sample_text}"')
+    print(f'Generated {len(result)} audio files from "{sample_text}"')
+
+def main(pdf_file: gr.File, output_filename: str):
+    start_time = time.time()
+    pdf_text = convert_pdf_to_text(pdf_file)
+    print(f'Processed PDF content in {time.time() - start_time:.4f} seconds')
+
+    pipe = QuantizedSpeechT5TTSPipe()
+    start_time = time.time()
+    audios = pipe.generate(pdf_text)
+    print(f'Generated {len(audios)} audio files in {time.time() - start_time:.4f} seconds')
+
+    zip_buffer = BytesIO()
+    with ZipFile(zip_buffer, mode='w') as zf:
+        for i, audio in enumerate(audios):
+            filename = f"{i}_{output_filename}.wav"
+            zf.writestr(filename, audio)
+    zip_buffer.seek(0)
+
+    return {'zip': zip_buffer}
+
+iface = gr.Interface(fn=main, 
+                      inputs="file", 
+                      outputs="binary",
+                      input_types=['pdf'],
+                      output_types=['download'])
+
+iface.launch()