Update app.py

app.py

@@ -1,25 +1,43 @@
 import subprocess
 
-subprocess.run(["pip", "install", "PyPDF2", "transformers", "bark", "gradio"])
+subprocess.run(["pip", "install", "PyPDF2", "transformers", "bark", "gradio","soundfile","PyMuPDF","numpy"])
 
 import PyPDF2
 from transformers import pipeline
 from bark import SAMPLE_RATE, generate_audio, preload_models
 import gradio as gr
+from IPython.display import Audio
+import os
+import io 
+import fitz
+import tempfile
+from PyPDF2 import PdfReader
+import numpy as np
+from tempfile import NamedTemporaryFile
+import soundfile as sf
 
-def summarize_and_convert_to_audio(pdf_file, abstract_page):
-    # Save the uploaded file
-    pdf_path = "/tmp/uploaded_file.pdf"
-    pdf_file.save(pdf_path)
+def readPDF(pdf_file_path):
+   if not pdf_file_path.endswith(".pdf"):
+        raise ValueError("Please upload a PDF file.")
+  
+   with open(pdf_file_path, 'rb') as file:
+       pdf_reader = file.read()
+   return pdf_reader
 
-    # Convert abstract_page to integer
-    abstract_page = int(abstract_page)
 
-    with open(pdf_path, 'rb') as file:
-        pdf_reader = PyPDF2.PdfReader(file)
-        
+def summarize_and_convert_to_audio(pdf_reader, page):
+   
+    temp_file = tempfile.NamedTemporaryFile(delete=False)
+    temp_file.write(pdf_reader)
+    temp_file_path = temp_file.name
+
+    # Use PyMuPDF to read the PDF content
+    pdf_document = fitz.open(temp_file_path)
+    
+    print(page)    
+
         # Get the abstract page text
-        abstract_page_text = pdf_reader.pages[abstract_page - 1].extract_text()
+    abstract_page_text = pdf_document[int(page) - 1].get_text()
 
     summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
     summary = summarizer(abstract_page_text, max_length=20, min_length=20)
@@ -29,16 +47,41 @@ def summarize_and_convert_to_audio(pdf_file, abstract_page):
     text = summary[0]['summary_text']
     audio_array = generate_audio(text)
 
-    return Audio(audio_array, rate=SAMPLE_RATE)
+    #save temporary file audio to use it in the second step 
+    
+    with NamedTemporaryFile(suffix=".wav", delete=False) as temp_wav_file:
+      wav_file_path = temp_wav_file.name
+      sf.write(wav_file_path, audio_array, SAMPLE_RATE)
+    return wav_file_path
+
+
+def read_and_speech(pdf_file,abstract_page):
+  print(pdf_file)
+  pdf_file_path= pdf_file.name
+  print(pdf_file_path)
+  page=abstract_page
+  reader=readPDF(pdf_file_path)
+  audio=summarize_and_convert_to_audio(reader,page)
+  return audio;
+    
+# Define app name, app description, and examples
+app_name = "From PDF to Speech"
+app_description = "Convert text from a PDF file to audio. Upload a PDF file. We accept only PDF files with abstracts."
+
 
 iface = gr.Interface(
-    fn=summarize_and_convert_to_audio,
+    fn=read_and_speech,
     inputs=[
-        gr.File("application/pdf", label="Upload PDF file"),
-        "text"
+        gr.File(file_types=["pdf"], label="Upload PDF file"),
+        gr.Textbox(label="Insert the page where the abstract is located")],
+    outputs=gr.Audio(type="filepath"),
+    title=app_name,
+    description=app_description,
+    examples=[
+        ["/content/drive/MyDrive/AAI/2312.04027.pdf",1],
+        ["/content/drive/MyDrive/AAI/2312.04542.pdf",1],
     ],
-    outputs="audio",
-    live=True
-)
+    allow_flagging="never"
+    )
 
 iface.launch()