Fix: Improve podcast analysis LLM prompt and skip bible generation in podcast mode

- Add pandas to requirements-podcast.txt for usage tracking
- Fix LLM prompt to return plain strings instead of objects for enhanced_ideas
- Add object-to-string normalization for LLM responses that return objects
- Skip bible generation in podcast-only mode (onboarding disabled)
- Skip alerts polling in AlertsBadge when in podcast-only demo mode

frontend/src/components/shared/AlertsBadge.tsx

@@ -222,6 +222,11 @@ const AlertsBadge: React.FC<AlertsBadgeProps> = ({ colorMode = 'light' }) => {
 
   // Poll for alerts
   useEffect(() => {
+    // Skip alerts polling entirely in podcast-only mode
+    if (isPodcastOnlyDemoMode()) {
+      return;
+    }
+
     if (!userId) return;
 
     // Delay initial fetch slightly to ensure auth token getter is installed