Update app.py

app.py

@@ -6,7 +6,7 @@ from sentence_transformers import SentenceTransformer
 import faiss
 from PyPDF2 import PdfReader
 from docx import Document
-from transformers import pipeline  # Hugging Face for summarization
+from transformers import pipeline
 
 # Initialize Sentence Transformer for embeddings
 model = SentenceTransformer('all-MiniLM-L6-v2')
@@ -52,11 +52,10 @@ def chunk_text(text, chunk_size=500, overlap=50):
 # Function to create embeddings and populate FAISS index
 def create_embeddings_and_store(chunks):
     global index
-    # Reset the FAISS index before adding new embeddings
     index = faiss.IndexFlatL2(dimension)
     for chunk in chunks:
         embedding = model.encode([chunk])
-        embedding = embedding.astype('float32')  # Ensure embedding is in correct format
+        embedding = embedding.astype('float32')
         index.add(embedding)
 
 # Function for summarizing the text before sending
@@ -65,9 +64,9 @@ def summarize_text(text):
     return summary[0]['summary_text']
 
 # Function to dynamically truncate context to fit the Groq API's token limit
-def truncate_context(context, max_tokens=4000):  # Adjust max_tokens based on Groq's limits
+def truncate_context(context, max_tokens=4000):
     if len(context) > max_tokens:
-        context = context[:max_tokens]  # Truncate context to fit within the token limit
+        context = context[:max_tokens]
     return context
 
 # Function to query Groq with context and question
@@ -78,11 +77,9 @@ def query_groq(question, context):
         if not context.strip():
             return "Error: No context available from the uploaded documents."
 
-        # Dynamically truncate context to fit within the token limit
-        max_context_tokens = 4000  # Groq's token limit for context
+        max_context_tokens = 4000
         context = truncate_context(context, max_tokens=max_context_tokens)
 
-        # Query Groq API with the truncated context
         chat_completion = client.chat.completions.create(
             messages=[{"role": "system", "content": "You are a helpful assistant. Use the context provided to answer the question."},
                       {"role": "assistant", "content": context},
@@ -101,131 +98,89 @@ def rag_pipeline(files, question, summarize_before_sending=False):
         if not files:
             return "Error: No files uploaded. Please upload at least one document."
 
-        # Process uploaded files
         texts = process_files(files)
         if not texts:
             return "Error: Could not extract text from the uploaded files."
 
-        # Combine all extracted text into a single context
         combined_text = " ".join(texts)
 
         if summarize_before_sending:
-            # Summarize the text to reduce token count
             combined_text = summarize_text(combined_text)
 
-        # Ensure the combined text is within Groq's token limit
-        max_text_size = 4000  # Adjust based on Groq's token limits
+        max_text_size = 4000
         combined_text = truncate_context(combined_text, max_tokens=max_text_size)
 
-        # Chunk and create embeddings
         chunks = chunk_text(combined_text)
         create_embeddings_and_store(chunks)
 
-        # Query Groq LLM with context and question
         answer = query_groq(question, combined_text)
         return answer
     except Exception as e:
         return f"Error: {str(e)}"
 
-# # Enhanced UI with modern and clean style
-# with gr.Blocks() as app:
-#     with gr.Row():
-#         # Left Column for instructions
-#         with gr.Column(scale=1, min_width=250):
-#             gr.Markdown("""
-#             <div style="background: linear-gradient(145deg, #6e7dff, #1c2b58); padding: 30px; border-radius: 12px; box-shadow: 0 5px 15px rgba(0, 0, 0, 0.1); font-family: 'Roboto', sans-serif;">
-#                 <h2 style="color: #fff; font-size: 32px; font-weight: bold;">DocAI: Document Assistant</h2>
-#                 <p style="color: #ddd; font-size: 18px;">Welcome to DocAI! Upload your documents and get intelligent answers based on their content.</p>
-#                 <p style="color: #ddd; font-size: 16px; line-height: 1.6;"><strong>Steps to use:</strong></p>
-#                 <ul style="color: #ddd; font-size: 16px; line-height: 1.6;">
-#                     <li>Upload your PDF or DOCX files.</li>
-#                     <li>Ask questions related to the document.</li>
-#                     <li>Enable "Summarize Before Sending" for a brief summary of the document.</li>
-#                     <li>Click "Submit" to get your answers.</li>
-#                 </ul>
-#                 <p style="color: #ddd; font-size: 16px; line-height: 1.6;">Upload multiple files and get answers based on their contents.</p>
-#             </div>
-#             """)
-
-#         # Right Column for the main application content
-#         with gr.Column(scale=2, min_width=600):
-#             gr.Markdown("""
-#             <div style="background: linear-gradient(135deg, #6e7dff, #1c2b58); padding: 20px; border-radius: 15px; box-shadow: 0 5px 15px rgba(0, 0, 0, 0.2); font-family: 'Roboto', sans-serif;">
-#                 <h2 style="color: #fff; font-size: 36px; font-weight: bold; text-align: center; letter-spacing: 2px; text-transform: uppercase;">
-#                     Ask Your Document
-#                 </h2>
-#                 <p style="color: #ddd; font-size: 18px; text-align: center; line-height: 1.6;">
-#                     Get intelligent answers based on the content of your uploaded documents. Just ask a question!
-#                 </p>
-#             </div>
-#             """)
-
-#             # File input
-#             file_input = gr.File(
-#                 label="Upload Documents (PDF/DOCX)", 
-#                 file_types=[".pdf", ".docx"], 
-#                 file_count="multiple", 
-#                 interactive=True
-#             )
-
-#             # Question input
-#             question_input = gr.Textbox(
-#                 label="Ask a question", 
-#                 placeholder="Type your question here...", 
-#                 interactive=True, 
-#                 lines=2, 
-#                 max_lines=4
-#             )
-
-#             # Summarize before sending checkbox
-#             summarize_before_input = gr.Checkbox(
-#                 label="Summarize Before Sending", 
-#                 value=False
-#             )
-
-#             # Output text box with enhanced styling
-#             output = gr.Textbox(
-#                 label="Answer from LLM", 
-#                 interactive=False, 
-#                 lines=4, 
-#                 max_lines=6
-#             )
-            
-#             # Submit button with icon and modern styling
-#             submit_button = gr.Button("Submit", icon="send")
-
-#             # Loading spinner
-#             with gr.Row():
-#                 with gr.Column(scale=1, min_width=250):
-#                     gr.Markdown("<div style='font-size: 14px; color: #555;'>Your answer will appear here...</div>")
-            
-#             # Apply the logic for the button to trigger the RAG pipeline
-#             submit_button.click(rag_pipeline, inputs=[file_input, question_input, summarize_before_input], outputs=output)
-
-# Launch the app
-# app.launch()
+# Enhanced UI with modern and clean style
 with gr.Blocks() as app:
     with gr.Row():
+        # Left Column for instructions
         with gr.Column(scale=1, min_width=250):
             gr.Markdown("""
-                <h2>Welcome to DocAI</h2>
-                <p>Upload your documents and get intelligent answers.</p>
+            <div style="background: linear-gradient(145deg, #6e7dff, #1c2b58); padding: 30px; border-radius: 12px; box-shadow: 0 5px 15px rgba(0, 0, 0, 0.1); font-family: 'Roboto', sans-serif;">
+                <h2 style="color: #fff; font-size: 32px; font-weight: bold;">DocAI: Document Assistant</h2>
+                <p style="color: #ddd; font-size: 18px;">Welcome to DocAI! Upload your documents and get intelligent answers based on their content.</p>
+                <p style="color: #ddd; font-size: 16px; line-height: 1.6;"><strong>Steps to use:</strong></p>
+                <ul style="color: #ddd; font-size: 16px; line-height: 1.6;">
+                    <li>Upload your PDF or DOCX files.</li>
+                    <li>Ask questions related to the document.</li>
+                    <li>Enable "Summarize Before Sending" for a brief summary of the document.</li>
+                    <li>Click "Submit" to get your answers.</li>
+                </ul>
+                <p style="color: #ddd; font-size: 16px; line-height: 1.6;">Upload multiple files and get answers based on their contents.</p>
+            </div>
             """)
 
+        # Right Column for the main application content
         with gr.Column(scale=2, min_width=600):
             gr.Markdown("""
-                <h2>Ask Your Document</h2>
-                <p>Get intelligent answers based on the content of your uploaded documents. Just ask a question!</p>
+            <div style="background: linear-gradient(135deg, #6e7dff, #1c2b58); padding: 20px; border-radius: 15px; box-shadow: 0 5px 15px rgba(0, 0, 0, 0.2); font-family: 'Roboto', sans-serif;">
+                <h2 style="color: #fff; font-size: 36px; font-weight: bold; text-align: center; letter-spacing: 2px; text-transform: uppercase;">
+                    Ask Your Document
+                </h2>
+                <p style="color: #ddd; font-size: 18px; text-align: center; line-height: 1.6;">
+                    Get intelligent answers based on the content of your uploaded documents. Just ask a question!
+                </p>
+            </div>
             """)
 
             # File input
-            file_input = gr.File(label="Upload Documents (PDF/DOCX)", file_types=[".pdf", ".docx"], file_count="multiple")
+            file_input = gr.File(
+                label="Upload Documents (PDF/DOCX)", 
+                file_types=[".pdf", ".docx"], 
+                file_count="multiple", 
+                interactive=True
+            )
 
             # Question input
-            question_input = gr.Textbox(label="Ask a question", placeholder="Type your question here...")
+            question_input = gr.Textbox(
+                label="Ask a question", 
+                placeholder="Type your question here...", 
+                interactive=True, 
+                lines=2, 
+                max_lines=4
+            )
+
+            # Summarize before sending checkbox
+            summarize_before_input = gr.Checkbox(
+                label="Summarize Before Sending", 
+                value=False
+            )
 
             # Output text box
-            output = gr.Textbox(label="Answer from LLM")
+            output = gr.Textbox(
+                label="Answer from LLM", 
+                interactive=False, 
+                lines=4, 
+                max_lines=6
+            )
 
             # Submit button
             submit_button = gr.Button("Submit", icon="send")