Follow up to T419466: Spike - Title normalization is not applied when saving pages to the reading list via the API

We need to create a maintenance script that applies normalization (space -> underscore) to existing reading list entry page titles for any pages that currently are saved with spaces, and then soft-delete any duplicates (with saved with the space form)

• Identify entries saved with spaces.
• Check if there is a duplicate entry saved with underscores.
  • If the entry (with spaces) is a duplicate, soft delete it
  • If no duplicates, then apply normalization (convert spaces to underscores)
• Adjust rl_size in the reading_list database table