Added code for file uploads

app.py

@@ -2,8 +2,10 @@ import gradio as gr
 import requests
 from newspaper import Article
 from transformers import pipeline
-import os
+import config
 import nltk
+import os
+import PyPDF2
 
 
 # Load summarization pipeline
@@ -56,14 +58,50 @@ def summarize_input(mixed_input):
         return f"**Title:** {title}\n\n**Author(s):** {author}\n\n**Published:** {pub_date}\n\n**📝 Summary** \n\n{summary}\n\n[🔗 Read more]({mixed_input})\n\n---"
     else:
         summary = generate_summary(mixed_input)
-        return f"## 📝 Summary \n\n{summary}\n\n**Original Text:**\n\n{mixed_input}\n\n---"
+        return f"## 📝 Summary \n\n{summary}\n\n📎 **Original Text:**\n\n{mixed_input}\n\n---"
+    
+# Function to summarize a file (PDF or TXT)
+def summarize_file(file):
+    try:
+        if file is None:  # Handle the case where no file is provided
+            return ""  # Return an empty string instead of an error message
+
+        text = ""
+        if file.name.endswith(".pdf"):
+            with open(file.name, "rb") as f:
+                reader = PyPDF2.PdfReader(f)
+                for page in reader.pages:
+                    text += page.extract_text() or ""
+        elif file.name.endswith(".txt"):
+            with open(file.name, "r", encoding="utf-8") as f:
+                text = f.read()
+        else:
+            return "❌ Unsupported file type."
+
+        if not text.strip():
+            return "❌ No text found in file."
+
+        summary = generate_summary(text)
+        original_text = text
+
+        # Combine the outputs into a single string
+        result = (
+            f"### 📝 Summary\n\n"
+            f"{summary}\n\n"
+            f"---\n\n"
+            f"📎 **Original Extracted Text:**\n\n{original_text}"
+        )
+        return result
+    except Exception as e:
+        return f"❌ Error processing file: {str(e)}"
+
+
 
 # Function to fetch top headlines from NewsAPI and summarize them
 def fetch_news():
     url = 'https://newsapi.org/v2/top-headlines'
-    api_key = os.environ.get("api_key")
     params = {
-        'apiKey': api_key,
+        'apiKey': config.api_key,
         'language': 'en',
         'sources': 'associated-press',
         'pageSize': 10
@@ -82,7 +120,7 @@ def fetch_news():
             pub_date = article.get("publishedAt", "Unknown")
             content = extract_full_content(article_url) or article.get("content") or article.get("description") or ""
             summary = generate_summary(content)
-            summaries.append(f"**{title}** \n\n**Author(s):** {author}\n\n**Published:** {pub_date}\n\n**Summary:** {summary}\n\n [🔗 Read more]({article_url})\n\n---")
+            summaries.append(f"**{title}** \n\n**Author(s):** {author}\n\n**Published:** {pub_date}\n\n**📝 Summary:** {summary}\n\n [🔗 Read more]({article_url})\n\n---")
 
         if not summaries:
             return "### No articles could be summarized."
@@ -101,29 +139,106 @@ def extract_full_content(url):
         return None
 
 # Gradio interface
-with gr.Blocks(theme=gr.themes.Base()) as demo:
+with gr.Blocks(theme=gr.themes.Default(font="Arial", font_mono="Courier New")) as demo:
+    # Header Section
     gr.Markdown("# 📰 Sum Up! Stay Informed, Instantly")
-    gr.Markdown(" ## A LLM based News Summarizer App")
+    gr.Markdown("## A LLM-based News Summarizer App")
+    gr.Markdown(
+        "Sum Up! condenses the latest headlines from trusted news sources into clear, concise, and easy-to-read summaries, so you can stay informed in seconds."
+    )
 
-    # Add a brief description
-    gr.Markdown("Sum Up! condenses the latest headlines from trusted news sources into clear, concise and easy-to-read summaries, so you can stay informed in seconds.")
+    # Input Section
+    gr.Markdown("---")  # Horizontal line for separation
+    with gr.Row():
+        # Left Column: Collapsible Sidebar for Latest News
+        with gr.Column(scale=1, min_width=300):
+            with gr.Accordion("📢 Fetch Latest News", open=False):
+                gr.Markdown("**Source: Associated Press**")
+                gr.Markdown(
+                    "Click the button below to fetch the top news articles and summarize them."
+                )
+                news_btn = gr.Button("🗞️ News Now", variant="primary", elem_id="news-now-btn")
+                gr.Markdown(
+                    "This feature fetches the latest headlines from the Associated Press and summarizes them for you."
+                )
+
+        # Right Column: Text Input and File Upload
+        with gr.Column(scale=2, min_width=400):
+            gr.Markdown("### Provide Your Input")
+            gr.Markdown("#### Enter Text or URL")
+            input_box = gr.Textbox(
+                label="Enter URL or Text",
+                placeholder="Paste a URL or text here...",
+                lines=5,
+            )
+            summarize_btn = gr.Button("📝 Summarize", variant="primary", elem_id="summarize-btn")
+
+            # Clear Button placed below the Summarize button
+            clear_btn = gr.Button("Clear", variant="secondary", elem_id="clear-btn")
+
+            gr.Markdown("#### Upload a File")
+            file_input = gr.File(
+                label="Upload a .pdf or .txt file", file_types=[".pdf", ".txt"]
+            )
+            gr.Markdown("**Note:** Only PDF and TXT files are supported.")
+
+    # Output Section
+    gr.Markdown("---")  # Horizontal line for separation
+    gr.Markdown("### View Results")
     with gr.Row():
         with gr.Column(scale=1):
-            gr.Markdown("### Top Stories - A Snapshot ")
-            gr.Markdown("**Source: Associated Press**")
-            gr.Markdown("Click the button below to fetch the latest news articles.")
-            news_btn = gr.Button("🗞️ News Now", variant="primary")
-        with gr.Column(scale=1):
-            input_box = gr.Textbox(label="Enter article text or URL", placeholder="Paste article text or link...")
-            summarize_btn = gr.Button("🔍 Summarize", variant="secondary")
-
-    # Output area for displaying results
-    output_area = gr.Markdown()  # Use a valid output component
+            gen_output = gr.Markdown()  # Use a valid output component
 
     # Link buttons to their respective functions
-    summarize_btn.click(fn=summarize_input, inputs=input_box, outputs=output_area)
-    news_btn.click(fn=fetch_news, inputs=[], outputs=output_area)
-
+    summarize_btn.click(fn=summarize_input, inputs=input_box, outputs=gen_output)
+    file_input.change(fn=summarize_file, inputs=file_input, outputs=gen_output)
+    news_btn.click(fn=fetch_news, inputs=[], outputs=gen_output)
+
+    # Clear button functionality
+    clear_btn.click(
+        fn=lambda: ("", None, ""),  # Clear all inputs and outputs
+        inputs=[],
+        outputs=[input_box, file_input, gen_output],
+    )
+
+# Ensure gen_output is properly reset
+gen_output = gr.Markdown(value="")  # Initialize with an empty value
+
+# Add custom CSS for better styling
+css = """
+#summarize-btn {
+    background-color: #4CAF50 !important; /* Green for Summarize */
+    color: white !important;
+    font-size: 16px !important;
+    padding: 10px 20px !important;
+    border-radius: 5px !important;
+    margin-top: 20px !important;
+    width: 100%;
+}
+
+#news-now-btn {
+    background-color: #0078D7 !important; /* Blue for News Now */
+    color: white !important;
+    font-size: 16px !important;
+    padding: 10px 20px !important;
+    border-radius: 5px !important;
+    margin-top: 20px !important;
+    width: 100%;
+}
+
+#clear-btn {
+    background-color: #d6d8db !important; /* Lighter Gray for Clear */
+    color: black !important;
+    font-size: 16px !important;
+    padding: 10px 20px !important;
+    border-radius: 5px !important;
+    margin-top: 20px !important;
+    width: 100%;
+}
+"""
+
+# Apply the custom CSS
+demo.css = css
 
 if __name__ == "__main__":
     demo.launch()

requirements.txt

@@ -5,4 +5,4 @@ torch
 newspaper3k
 requests
 gradio
-lxml_html_clean
+PyPDF2