Fix HF Space deployment - lazy load chatbot

Dockerfile

@@ -2,6 +2,11 @@ FROM python:3.11-slim
 
 WORKDIR /app
 
+# Install system dependencies
+RUN apt-get update && apt-get install -y \
+    build-essential \
+    && rm -rf /var/lib/apt/lists/*
+
 # Copy requirements and install dependencies
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
@@ -13,4 +18,5 @@ COPY . .
 EXPOSE 7860
 
 # Run the application with gunicorn
-CMD ["gunicorn", "--bind", "0.0.0.0:7860", "--workers", "4", "--timeout", "120", "app:app"]
+# Reduced workers to 2 and increased timeout for model loading
+CMD ["gunicorn", "--bind", "0.0.0.0:7860", "--workers", "2", "--timeout", "300", "--preload", "app:app"]

app.py

@@ -10,7 +10,6 @@ load_dotenv()
 sys.path.insert(0, os.path.join(os.path.dirname(__file__), 'src'))
 from recommendation_engine import RecommendationEngine
 from recommendation_explainer import RecommendationExplainer
-from rag import RAGChatbot
 
 app = Flask(__name__)
 app.secret_key = os.environ.get('SECRET_KEY', 'dev-secret-key-change-in-production')
@@ -21,14 +20,29 @@ recommendation_explainer = RecommendationExplainer()
 
 # Initialize RAG chatbot (lazy initialization on first request)
 rag_chatbot = None
+chatbot_initialized = False
+chatbot_error = None
 
 def get_chatbot():
     """Get or initialize the RAG chatbot"""
-    global rag_chatbot
-    if rag_chatbot is None:
+    global rag_chatbot, chatbot_initialized, chatbot_error
+
+    if chatbot_initialized:
+        return rag_chatbot
+
+    try:
+        print("🤖 Initializing RAG chatbot...")
+        from rag import RAGChatbot
         rag_chatbot = RAGChatbot()
         rag_chatbot.initialize()
-    return rag_chatbot
+        chatbot_initialized = True
+        print("✅ RAG chatbot initialized successfully")
+        return rag_chatbot
+    except Exception as e:
+        chatbot_error = str(e)
+        chatbot_initialized = True  # Mark as initialized to avoid retrying
+        print(f"❌ Failed to initialize chatbot: {e}")
+        return None
 
 # Custom Jinja filter for simple markdown rendering
 import re
@@ -130,6 +144,13 @@ def chat():
 
         # Get chatbot and generate response
         chatbot = get_chatbot()
+
+        if chatbot is None:
+            return jsonify({
+                "response": "Sorry, the chatbot is currently unavailable. Please try again later or contact support.",
+                "status": "success"
+            })
+
         response = chatbot.chat(query)
 
         return jsonify({
@@ -137,11 +158,13 @@ def chat():
             "status": "success"
         })
     except Exception as e:
-        print(f"Error in chat endpoint: {e}")
+        print(f"❌ Error in chat endpoint: {e}")
+        import traceback
+        traceback.print_exc()
         return jsonify({
-            "error": "Sorry, I encountered an error processing your question.",
-            "status": "error"
-        }), 500
+            "response": "Sorry, I encountered an error processing your question. Please try again.",
+            "status": "success"
+        })
 
 if __name__ == '__main__':
     app.run(debug=True, host='0.0.0.0', port=5555)

requirements.txt

@@ -9,3 +9,9 @@ pandas==2.3.0
 plotly==5.24.1
 openai>=1.0.0
 gunicorn==23.0.0
+requests>=2.32.0
+python-frontmatter>=1.1.0
+sentence-transformers>=2.2.0
+tqdm>=4.65.0
+numpy>=1.24.0
+minsearch>=0.0.7

templates/base.html

@@ -335,21 +335,28 @@
                     headers: {
                         'Content-Type': 'application/json',
                     },
-                    body: JSON.stringify({ message: message })
+                    body: JSON.stringify({ message: message }),
+                    signal: AbortSignal.timeout(60000) // 60 second timeout
                 });
 
                 const data = await response.json();
                 removeTyping();
 
-                if (data.status === 'success') {
+                if (data.status === 'success' && data.response) {
                     addMessage(data.response);
+                } else if (data.error) {
+                    addMessage(data.error);
                 } else {
                     addMessage('Sorry, I encountered an error. Please try again.');
                 }
             } catch (error) {
                 console.error('Chat error:', error);
                 removeTyping();
-                addMessage('Sorry, I encountered an error. Please try again.');
+                if (error.name === 'TimeoutError') {
+                    addMessage('Sorry, the request timed out. The chatbot may be initializing. Please try again in a moment.');
+                } else {
+                    addMessage('Sorry, I encountered an error. Please try again later.');
+                }
             } finally {
                 // Re-enable input
                 chatbotInput.disabled = false;