Fix cleanup batching + reduce DB load on high-traffic sites

[gsusI]

Fixes #9.

Changes

• Fix cleanup threshold calculation to be relative to "now" (was hard-coded), so old rows are actually purged.
• Re-enable iterative batch cleanup scheduling until the table is within the retention window.
• Remove cron-time output (no echo).
• Bound rsp_get_most_visited() aggregation to the same retention window (RSP_CLEANUP_THREASHOLD_TIME) to keep the query bounded if cron is delayed.
• Use a single-statement INSERT ... ON DUPLICATE KEY UPDATE for visit recording to reduce DB queries per page view and keep updates atomic.
• Add an index on last_visit in the table schema to speed cleanup / range queries (via dbDelta on activation).
• Add an admin notice + CTA to add the last_visit index on demand (so high-traffic sites can run it off-peak).
• Bump package version to 1.1.1.

Notes

• The DB update notice is shown on the Plugins screen and the plugin settings page until an index on last_visit exists.
• The CTA runs an ALTER TABLE ... ADD INDEX ... which can take time and lock writes on very large tables.

Related: #8