Fix performance for filter today events function

[hoangqwe159]

Fix Performance Regression in Recurring Event Timezone Conversion

Problen and fix demo

https://drive.google.com/file/d/1aLZ7Pyqvd9ilqmWYt0st8t-9bQjmfUtO/view?usp=sharing

Summary

This PR fixes a significant performance regression introduced in the recent timezone handling refactor for recurring events. The previous implementation was performing timezone conversions on every recurrence instance, causing O(n × r) complexity instead of O(n), where r = number of recurrences per event.

Problem

The recent changes to getRecurrencesForDate and filterTodayEvents introduced several performance issues:

1. Redundant timezone conversions: Each recurrence instance was being converted individually instead of converting the base event once
2. Expanded search window: Changed from 1-day to 2-day search window, generating more recurrences unnecessarily
3. Increased computational overhead: For events with many recurrences (e.g., hourly meetings), this created significant bottlenecks

Example Impact

• Before fix: Event with 100 daily recurrences → 100 Intl.DateTimeFormat calls
• After fix: Event with 100 daily recurrences → 1 Intl.DateTimeFormat call

Changes

getRecurrencesForDate Function

• Removed timeZone parameter from function signature
• Removed timezone conversion logic that was being applied to each recurrence
• Restored original 1-day search window: between(today, addDays(today, 1), true)
• Kept the convertRRuleDateToDate utility for proper RRule date handling

filterTodayEvents Function

• Restored pre-conversion pattern: convert each event's timezone ONCE before generating recurrences
• Changed from getRecurrencesForDate(events[i], today, timeZone) to:

  const event = convertEventTimeZone(events[i], timeZone);
  for (const rec of getRecurrencesForDate(event, today)) {

Testing

• Verified recurring events display correctly across different timezones
• Confirmed performance improvement with high-frequency recurring events
• Tested edge cases: all-day events, multi-day events, single events
• Validated RRule date conversion still works correctly

Performance Metrics

• Time complexity: Restored from O(n × r) to O(n)
• Expected improvement: Up to 100x faster for events with 100+ recurrences

Breaking Changes

None - this is a performance optimization that maintains the same external behavior.