Commit Graph

2 Commits

Author SHA1 Message Date
a7f3b215e4 perf: optimize loading time — paginated API, slim payloads, local category filtering
Backend: Rewrote EventListAPI to query per-type with DB-level LIMIT
instead of loading all 734 events into memory. Added slim serializer
(32KB vs 154KB). Added DB indexes on event_type_id and pincode.

Frontend: Category chips now filter locally from _allEvents (instant,
no API call). Top Events and category sections always show all types
regardless of selected category. Added TTL caching for event types
(30min) and events (5min). Reduced API timeout from 30s to 10s.
Added memCacheHeight to all CachedNetworkImage widgets. Batched
setState calls from 5 to 2 during startup. Cached _eventDates getter.

Switched baseUrl to em.eventifyplus.com (Django via Nginx+SSL).
Added initialEvent param to LearnMoreScreen for instant detail views.
Resolved relative media URLs for category icons.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-30 10:05:23 +05:30
0c4e62d00e perf: add memCacheWidth/memCacheHeight to all thumbnail images
All CachedNetworkImage instances in list/card contexts now decode at
2x rendered size instead of full resolution. A 3000x2000 event photo
previously decoded to ~24MB in GPU memory even when shown at 96px —
now decodes to <1MB.

Affected screens (16 CachedNetworkImage instances total):
- home_screen.dart: hero (800w), top card (300w), stacked (192w),
  horizontal (440x360), full-width (800x400), search (112x112),
  filter sheet (160x160), type icons (112x112)
- home_desktop_screen.dart: mini (128x128), grid (600x280), horiz (600x296)
- calendar_screen.dart: event card (400x300)
- profile_screen.dart: avatar (size*2), event tile (120x120)

learn_more_screen.dart intentionally unchanged — full-res for detail view.

Estimated memory reduction: ~500MB → ~30MB for a typical home screen.
2026-03-20 22:26:52 +05:30