Files
step3/scripts
hz 4251885a73 Redesign dashboard top-4 stats as separate cards
Frontend half of the dashboard stat-card redesign, matching the
provided mockup. The 4 metrics used to be plain rows sharing one
bordered box (_add_metric -> a single metric_grid inside one _card).
Replaced with 4 independently-_card-wrapped panels in a responsive
GridContainer (4 cols / 2 compact, same threshold _update_responsive
already used).

Each card now has: a colored icon badge, title + info glyph, a big
value, a colored trend/status line, a small visual, and a footer
caption:
- Question Bank Rating / CCS Rating: a new minimal DashboardSparkline
  (dashboard_sparkline.gd) fed from the current rating and the
  week-ago snapshot (reconstructed client-side from the delta the API
  already returns - no separate history endpoint needed). Trend line
  and caption come from the new delta/percentile fields; both degrade
  to a plain single-point sparkline and a rated-count caption when no
  snapshot exists yet.
- Exam Readiness: a new DashboardSegmentBar (dashboard_segment_bar.gd -
  nothing like a segmented pill progress indicator existed anywhere in
  the codebase) instead of a sparkline, colored by the High/Medium/Low
  label; shows a neutral "complete a simulation" state when the user
  has none yet rather than a fabricated readiness.
- Average Score: reuses the existing ProgressBar/_style_progress,
  relabeled from "First-Attempt Accuracy".

Dropped the old "Study Time" card (not part of the mockup - the
underlying data still drives the weekly activity chart below, it just
no longer gets a dedicated top-line number) and the now-fully-unused
_format_duration.

Verified headlessly: instantiated the real screen against fake
DashboardService data, confirmed every card's value/trend/caption text
and the segment-bar/progress-bar values match the mockup exactly, and
rendered actual screenshots of both the populated and "no data yet"
fallback states.
2026-07-26 00:43:39 -05:00
..