Update app.py

app.py

@@ -1,30 +1,72 @@
 import torch
 import gradio as gr
 from transformers import pipeline
+import logging
+
+# Set up logging
+logging.basicConfig(level=logging.INFO)
 
 # Use a pipeline as a high-level helper
 device = 0 if torch.cuda.is_available() else -1
-text_summary = pipeline("summarization", model="Falconsai/text_summarization", device=device, torch_dtype=torch.bfloat16)
+text_summary = pipeline("summarization", model="facebook/bart-large-cnn", device=device, torch_dtype=torch.bfloat16)
+
+# Function for summarization with enhancements
+def summary(input, summary_type="medium"):
+    # Check for empty input
+    if not input.strip():
+        return "Error: Please provide some text to summarize."
 
-def summary(input):
-    # Calculate the number of tokens based on input length
+    # Calculate input length
     input_length = len(input.split())
-    max_output_tokens = max(20, input_length // 2)  # Ensure output is less than half of the input
-    min_output_tokens = max(10, input_length // 4)  # Ensure a meaningful summary
+    logging.info(f"Input length: {input_length} words")
+
+    # Handle input that's too short
+    if input_length < 10:
+        return "Error: Input is too short. Please provide at least 10 words."
 
-    # Generate summary with dynamic token limits
+    # Handle input that's too long for the model
+    if input_length > 512:
+        return "Warning: Input exceeds the model's limit of 512 tokens. Please shorten the input text."
+
+    # Adjust max/min lengths based on the summary type
+    if summary_type == "short":
+        max_output_tokens = max(10, input_length // 4)
+    elif summary_type == "medium":
+        max_output_tokens = max(20, input_length // 2)
+    elif summary_type == "long":
+        max_output_tokens = max(30, (3 * input_length) // 4)
+    min_output_tokens = max(10, input_length // 6)
+
+    # Generate summary
     output = text_summary(input, max_length=max_output_tokens, min_length=min_output_tokens, truncation=True)
     return output[0]['summary_text']
 
+# Gradio interface
+def save_output(summary_text):
+    """Save the summarized text to a file."""
+    with open("summary_output.txt", "w") as file:
+        file.write(summary_text)
+    return "Summary saved to 'summary_output.txt'."
+
 gr.close_all()
 
 # Create the Gradio interface
 demo = gr.Interface(
     fn=summary,
-    inputs=[gr.Textbox(label="INPUT THE PASSAGE TO SUMMARIZE", lines=10)],
-    outputs=[gr.Textbox(label="SUMMARIZED TEXT", lines=4)],
+    inputs=[
+        gr.Textbox(label="INPUT THE PASSAGE TO SUMMARIZE", lines=15, placeholder="Paste your text here."),
+        gr.Dropdown(["short", "medium", "long"], label="SUMMARY LENGTH", value="medium")
+    ],
+    outputs=[
+        gr.Textbox(label="SUMMARIZED TEXT", lines=10, placeholder="Your summarized text will appear here."),
+        gr.Button("Download Summary")
+    ],
     title="PAVISHINI @ GenAI Project 1: Text Summarizer",
-    description="This application is used to summarize the text"
+    description=(
+        "This application summarizes input text. "
+        "The output length can be short, medium, or long based on your selection."
+    ),
+    live=True
 )
 
 demo.launch()