Add logging to Python module

[moksh-m9u]

The backend (app/) currently lacks structured logging and error handling,
making it difficult to trace failures across the AI query pipeline.

Proposed Changes

• Add Python's built-in logging module across the backend pipeline
• Wrap critical operations in try/except blocks with errors logged at
  appropriate levels (DEBUG, WARNING, ERROR)

Target Areas

• app/services/ — LLM calls, SQL generation, and self-correcting SQL retries
• app/handlers/ — Business logic execution
• app/api/ — Route-level request/response logging
• app/db/ — Database connection and session errors

Benefits

• Easier to trace where failures occur in the NL → SQL → execution pipeline
• No new dependencies — stdlib logging only
• Improves long-term maintainability of the backend

I'd like to work on this if you're open to it.