8000 Claude/studying ahead plan by jmahmood · Pull Request #24 · jmahmood/CardBrick · GitHub
[go: up one dir, main page]

Skip to content

Claude/studying ahead plan - #24

Merged
jmahmood merged 6 commits into
masterfrom
claude/studying-ahead-plan-7cmij2
Jul 8, 2026
Merged

jmahmood merged 6 commits into
masterfrom
claude/studying-ahead-plan-7cmij2

Conversation

@jmahmood
@jmahmood jmahmood commented Jul 7, 2026
Copy link
Copy Markdown
Owner

No description provided.

claude added 6 commits July 7, 2026 17:17
A profile migrated from the pre-sprint model kept its 50-card/15-min
session limits, so the whole daily goal fit in one 'sprint', and the
start screen promised the full goal (0/50) on a freshly imported deck
that could only supply daily_new_cards cards — then dead-ended after
them.

- Schema v6 re-baselines profiles still on the old 50/15 defaults to
  sprint scale (20 cards / 10 min); custom values are untouched.
- sprint_status now derives the plan from what today can actually
  supply: goal_today = done + min(goal left, fresh cards available),
  so day one of a fresh deck honestly reads '1 sprint, 0/10 cards'.
- Cards already answered today can't advance the distinct-card goal,
  so once only repeats remain the day completes instead of showing
  'sprints to go' forever.
- Bonus sprints may pull extra new cards beyond the daily pacing cap
  (one optional sprint at a time), so finishing the paced cards on a
  fresh deck offers 'bonus sprint: 20 cards' instead of 'come back
  tomorrow'.

Verified by driving the exact reported scenario headlessly: a
v5-migrated 50/15 profile with two freshly imported decks now shows
'Last sprint of the day! 0/10 cards', completes the 10, and offers a
bonus sprint of new cards.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BpqXQKjiwd3ogsmumtDe2j
Deletes the database (decks, review history, sessions, profiles) and
imported media after a timestamped backup, keeping settings.json and
input_mapping.json so display/controller calibration survives. The
next start recreates an empty database with a default profile.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BpqXQKjiwd3ogsmumtDe2j
'Last sprint of the day!' greeted a day where nothing had been studied
yet. The countdown phrasing now only appears once earlier sprints
happened; an untouched day reads 'Just one sprint today!' / 'N sprints
today', and a supply-limited day (fresh deck paced by daily_new_cards)
says why: 'more unlock tomorrow'.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BpqXQKjiwd3ogsmumtDe2j
The fixed 10-new-cards/day drip contradicted the goal model: a fresh
deck's whole day collapsed to a single 10-card sprint no matter the
goal. By default new cards now fill whatever the goal has left after
due reviews — day one of a fresh deck is a full goal-sized day of
sprints, and as the review load grows it crowds new cards out
automatically, so the goal itself is the pacing. A non-zero
daily_new_cards still imposes a stricter fixed cap.

Schema v7 switches profiles on the old default (10) to auto; custom
caps are kept.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BpqXQKjiwd3ogsmumtDe2j
The deck-name + tags header on the left and the sprint/remaining
label on the right shared one row with no width check, so a long deck
name or several tags (e.g. a vocab card with 'isla::Gramatica
isla::Directores') ran straight into 'Sprint 1/8 · 8 left'. The left
side is now measured and truncated with an ellipsis to whatever space
is actually left after the right side, so the two can never collide.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BpqXQKjiwd3ogsmumtDe2j
Two changes:

1. Fix: the Categories/Decks assignment screen in Parent Mode blitted
   each row at a fixed x with no width check, so a long tag name ran
   off the right edge of the screen with no wrap or ellipsis. The
   review header's start/goal labels had the same latent issue (a long
   joined tag/deck list centers past both edges since _center doesn't
   clip). All now go through the new _truncate_to_width helper.

2. Feature: a child-facing CATEGORY_SELECT screen, mirroring the
   existing DECK_SELECT picker — 'Choose a Topic' offers 'All assigned
   categories' plus one entry per parent-assigned tag, for just this
   sitting. Picking a tag is pure session-level filtering, the same
   non-destructive idea as Anki's filtered decks: no card's tags or
   deck ever change, so the tag is simply back in the pool next time
   'All assigned categories' (or nothing further) is picked. Skipped
   automatically when 0-1 categories are assigned, same threshold as
   the deck picker. StudySession gained a category_filter override
   parameter alongside the existing deck_filter one, and the summary
   screen's 'next sprint now' / bonus-sprint offer both preserve the
   chosen topic.

Covered by tests/test_category_select_session.py (override, fallback,
combining with deck_filter, non-destructiveness, session-row
recording) and tests/test_truncate_to_width.py (including a tiny-width
edge case the first version of the helper got wrong: appending an
ellipsis that itself didn't fit). Verified visually by rendering the
Categories screen and the new picker with deliberately overflow-length
tags via a deterministic poll()-scripted harness.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BpqXQKjiwd3ogsmumtDe2j
@jmahmood
jmahmood merged commit 63affef into master Jul 8, 2026
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants

0