Add comprehensive logging throughout the application

[MohammedAlkindi]

Summary

This PR adds structured logging across the backend and frontend to improve observability and debugging capabilities. A new logging configuration module centralizes log setup, and logging statements are added at key points in the recommendation pipeline and HTTP request handling.

Key Changes

Backend Logging Infrastructure:

• Created app/logging_config.py with centralized logging configuration that respects LOG_LEVEL environment variable (defaults to INFO)
• Suppresses noisy uvicorn access logs in DEBUG mode to avoid duplication
• Uses ISO 8601 timestamp format for consistency

API & Request Logging:

• Added HTTP middleware in main.py to log all requests with method, path, status code, and duration
• Enhanced /recommend endpoint with detailed logging:
  • Info-level logs for incoming requests with all preference parameters
  • Debug-level logs for liked/disliked/avoided titles
  • Error handling with try-catch blocks that log failures and return appropriate HTTP 500 errors
  • Info-level logs for successful responses with result count and top score

Scoring Engine Logging:

• Added debug-level logging in score_item() to show per-item scoring details
• Added debug-level logging in rank_recommendations() to show ranking results and top scores

Catalog Service Logging:

• Added debug-level log when loading catalog
• Added info-level log showing total items loaded

Frontend Logging:

• Implemented structured logging utility with info/debug/warn/error methods
• Added curio-debug localStorage flag to enable/disable debug output
• Logs API request payloads and response details with timing information
• Logs errors with full error objects for better debugging

Implementation Details

• All backend loggers use __name__ for proper module identification
• Logging uses lazy string formatting (%) to avoid unnecessary string construction at lower log levels
• Frontend logs are prefixed with [Curio] or [Curio:debug] for easy identification
• Error logging includes exc_info=True for full stack traces
• Request timing uses time.perf_counter() for high-resolution measurements

https://claude.ai/code/session_016AW7kKjJdbSnspDi1fJ3q7