Update app.py

app.py

@@ -34,15 +34,15 @@ flan_t5_tokenizer = T5Tokenizer.from_pretrained(flan_t5_model_id)
 flan_t5_model = T5ForConditionalGeneration.from_pretrained(flan_t5_model_id)
 
 # Function to transcribe audio files
-def transcribe_audio(file_path):
-    result = whisper_pipe(file_path)
+def transcribe_audio(file):
+    result = whisper_pipe(file)
     return result['text']
 
 # Function to extract text and questions from PDF
-def extract_text_from_pdf(pdf_path):
+def extract_text_from_pdf(pdf_file):
     text = ""
     questions = []
-    with pdfplumber.open(pdf_path) as pdf:
+    with pdfplumber.open(pdf_file) as pdf:
         for page in pdf.pages:
             page_text = page.extract_text()
             if page_text:
@@ -98,17 +98,12 @@ def save_responses_to_pdf(responses, output_pdf_path):
     )
 
     content = []
-    # Add heading
-    heading = Paragraph("<b>FillUp by Umar Majeed</b>", styles['Title'])
-    content.append(heading)
-    content.append(Spacer(1, 12))
-
     for index, response in enumerate(responses, start=1):
         # Add the response number and content
-        file_heading = Paragraph(f"<b>File {index}:</b>", styles['Heading2'])
+        heading = Paragraph(f"<b>File {index}:</b>", styles['Heading2'])
         response_text = Paragraph(response.replace("\n", "<br/>"), response_style)
 
-        content.append(file_heading)
+        content.append(heading)
         content.append(Spacer(1, 6))  # Space between heading and response
         content.append(response_text)
         content.append(Spacer(1, 18))  # Space between responses
@@ -131,18 +126,26 @@ def process_files(audio_files, pdf_file):
     output_pdf_path = "output.pdf"
     save_responses_to_pdf(responses, output_pdf_path)
     
-    return output_pdf_path
+    # Return the PDF path and the generated responses
+    return output_pdf_path, "\n\n".join(responses)
 
 # Gradio interface definition
 interface = gr.Interface(
     fn=process_files,
     inputs=[
-        gr.Files(label="Upload Audio Files", type="file"),
-        gr.File(label="Upload PDF File", type="file")
+        gr.Files(label="Upload Audio Dataset"),
+        gr.File(label="Upload PDF File with Questions")
+    ],
+    outputs=[
+        gr.File(label="Download Output PDF"),
+        gr.Textbox(label="Generated Responses", lines=20, placeholder="The responses will be shown here...")
     ],
-    outputs=gr.File(label="Download Output PDF"),
     title="FillUp by Umar Majeed",
-    description="Upload audio files and a PDF file. The application will transcribe the audio, extract questions from the PDF, and generate a response PDF."
+    description="""This is a beta version of FillUp, an application designed to auto-fill predefined forms using call data. 
+    Upload the audio files from which you want to extract text and a PDF form that contains the questions to be answered. 
+    At the end, you will receive a PDF file with the responses.
+
+    For reference, you can download a sample form from [https://drive.google.com/drive/folders/13LolIqxufzysqNoGMfuCAvpA9AkbRfL7?usp=drive_link]. Use this dummy data to understand how the model works."""
 )
 
 # Launch the interface