Upload app.py

app.py

@@ -83,9 +83,13 @@ class CurriculumChatbot:
             # Create QA prompt template
             qa_template = """<|begin_of_text|><|start_header_id|>system<|end_header_id|>
 
-You are a helpful AI programming tutor. Answer questions about programming concepts clearly and educationally. If the question is about curriculum content, use the provided context. If not, provide a general programming answer.
+You are a helpful AI programming tutor. You MUST ALWAYS provide a clear, educational answer to every question. Never say you cannot answer or that you don't know.
 
-Make sure to check the curriculum content and answer the question based on the curriculum content if possible. The RAG is not perfect, so if the question is not related to that slide, change it to be related to question.
+If the question is about curriculum content, use the provided context to give a detailed, educational explanation. If the curriculum content doesn't perfectly match the question, adapt your answer to be relevant while using the curriculum information.
+
+If the question is not covered in the curriculum, provide a comprehensive general programming answer based on your knowledge.
+
+Always be educational, clear, and helpful.
 
 <|eot_id|><|start_header_id|>user<|end_header_id|>
 
@@ -177,7 +181,7 @@ Which slide is most relevant? Return only: "filename.pdf - Page X"
             # Calculate relevance score based on similarity
             curriculum_relevance_score = len([r for r in results if r.page_content.strip()])
         
-        # Generate LLM answer
+        # ALWAYS generate LLM answer (never fallback to raw text)
         if self.qa_chain:
             try:
                 if curriculum_relevance_score > 0:
@@ -201,81 +205,74 @@ Which slide is most relevant? Return only: "filename.pdf - Page X"
                 
             except Exception as e:
                 print(f"Error generating answer: {e}")
+                # Even if LLM fails, try to provide a helpful response
                 if curriculum_relevance_score > 0:
-                    answer = f"Based on the curriculum content:\n\n{results[0].page_content}"
+                    answer = f"Based on the curriculum content, here's what I found:\n\n{results[0].page_content}\n\n*Note: I'm having trouble generating a custom answer right now, but here's the relevant curriculum content.*"
                 else:
-                    answer = "I'm sorry, I couldn't generate an answer at the moment."
+                    answer = "I'm sorry, I couldn't generate an answer at the moment. Please try rephrasing your question."
         else:
-            # Fallback without LLM
+            # If no LLM available, still provide helpful response
             if curriculum_relevance_score > 0:
-                answer = f"Most relevant content from the curriculum:\n\n{results[0].page_content}"
+                answer = f"Based on the curriculum content:\n\n{results[0].page_content}\n\n*Note: AI generation is not available, but here's the relevant curriculum content.*"
             else:
-                answer = "No relevant content found in the curriculum."
+                answer = "I couldn't find relevant content in the curriculum for this question. Please try rephrasing or ask about a different programming topic."
         
-        # Get recommended slide
-        recommended_slide = None
-        if curriculum_relevance_score > 0 and self.slide_selection_chain:
-            try:
-                available_slides = self.get_available_slides_text()
-                slide_response = self.slide_selection_chain.run(
-                    question=query, 
-                    available_slides=available_slides
-                )
+        # Get the most relevant slide and its neighboring pages
+        relevant_slides = []
+        if curriculum_relevance_score > 0:
+            # Get the most relevant result
+            best_result = results[0]
+            filename = best_result.metadata["filename"]
+            page_number = best_result.metadata["page_number"]
+            
+            # Get the specific PDF and its pages
+            if filename in self.pdf_files:
+                pdf_path = self.pdf_files[filename]
+                doc = fitz.open(pdf_path)
+                total_pages = len(doc)
+                doc.close()
                 
-                # Extract filename and page from response
-                slide_response = slide_response.strip()
-                if "<|eot_id|>" in slide_response:
-                    slide_response = slide_response.split("<|eot_id|>")[-1].strip()
+                # Get the target page and neighboring pages (2 before, 2 after)
+                start_page = max(1, page_number - 2)
+                end_page = min(total_pages, page_number + 2)
                 
-                # Parse the response to get filename and page
-                match = re.search(r'(.+\.pdf)\s*-\s*Page\s*(\d+)', slide_response)
-                if match:
-                    filename = match.group(1)
-                    page_num = int(match.group(2))
-                    if filename in self.pdf_files:
-                        recommended_slide = self.get_pdf_page_image(self.pdf_files[filename], page_num)
-                        recommended_label = f"{filename} - Page {page_num}"
-                    else:
-                        # Fallback to most relevant result
-                        recommended_slide = self.get_pdf_page_image(
-                            self.pdf_files[results[0].metadata["filename"]], 
-                            results[0].metadata["page_number"]
-                        )
-                        recommended_label = f"{results[0].metadata['filename']} - Page {results[0].metadata['page_number']}"
-                else:
-                    # Fallback to most relevant result
-                    recommended_slide = self.get_pdf_page_image(
-                        self.pdf_files[results[0].metadata["filename"]], 
-                        results[0].metadata["page_number"]
-                    )
-                    recommended_label = f"{results[0].metadata['filename']} - Page {results[0].metadata['page_number']}"
-            except Exception as e:
-                print(f"Error selecting slide: {e}")
-                # Fallback to most relevant result
-                if results:
-                    recommended_slide = self.get_pdf_page_image(
-                        self.pdf_files[results[0].metadata["filename"]], 
-                        results[0].metadata["page_number"]
-                    )
-                    recommended_label = f"{results[0].metadata['filename']} - Page {results[0].metadata['page_number']}"
-        
-        # Get all slides for navigation
-        all_slides = self.get_all_slides()
+                for page_num in range(start_page, end_page + 1):
+                    img = self.get_pdf_page_image(pdf_path, page_num)
+                    if img:
+                        if page_num == page_number:
+                            # Highlight the most relevant page
+                            label = f"📌 {filename} - Page {page_num} (Most Relevant)"
+                        else:
+                            label = f"{filename} - Page {page_num}"
+                        relevant_slides.append((img, label))
+                
+                recommended_slide = relevant_slides[0][0] if relevant_slides else None
+                recommended_label = relevant_slides[0][1] if relevant_slides else None
+            else:
+                # Fallback if filename not found
+                recommended_slide = None
+                recommended_label = None
+        else:
+            # If no curriculum content, show a few slides from different PDFs
+            relevant_slides = []
+            for filename, pages in list(self.pdf_pages.items())[:3]:  # Show first 3 PDFs
+                for page_num in list(pages.keys())[:2]:  # Show first 2 pages of each
+                    img = self.get_pdf_page_image(self.pdf_files[filename], page_num)
+                    if img:
+                        relevant_slides.append((img, f"{filename} - Page {page_num}"))
+            recommended_slide = relevant_slides[0][0] if relevant_slides else None
+            recommended_label = relevant_slides[0][1] if relevant_slides else None
         
-        return answer, recommended_slide, recommended_label if 'recommended_label' in locals() else None, all_slides
+        return answer, recommended_slide, recommended_label, relevant_slides
 
 # --- Gradio UI ---
 chatbot = CurriculumChatbot()
 
 def gradio_chat(query):
-    answer, recommended_slide, recommended_label, all_slides = chatbot.chat(query)
+    answer, recommended_slide, recommended_label, relevant_slides = chatbot.chat(query)
     
-    # Create gallery with recommended slide first, then all slides
-    if recommended_slide and recommended_label:
-        gallery_items = [(recommended_slide, f"📌 {recommended_label} (Recommended)")]
-        gallery_items.extend(all_slides)
-    else:
-        gallery_items = all_slides
+    # Use the relevant slides (specific PDF with neighboring pages)
+    gallery_items = relevant_slides if relevant_slides else []
     
     return answer, gallery_items