Repository navigation
feat(studio): experimental FSRS-4.5 scheduler for flashcards (off by default) - #20
Conversation
…default) Applies the recommendation from the real-log evaluation without changing the default: the Plan tab offers Leitner (current, default) or FSRS-4.5 (experimental). FSRS rebuilds each card's memory from its own review log (no new stored state; switching back is always possible) and schedules the next review when predicted recall reaches 90 % (interval = stability, 1–365 days). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
ⓘ Qodo reviews are paused because the subscription is no longer active. Ask your workspace admin to reactivate the subscription to resume reviews. Manage billing |
|
Important
This repository does not receive automatic reviews because it has fewer than 10 stars. ⚙️ Run configuration
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: a6b1649f97
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| import type { AskCourseAnswer } from "../academic-workspace/types"; | ||
| import { addDays, applyReview, type DayKey, type ReviewOutcome, type ReviewState } from "../daily-english/spaced-review"; | ||
| import type { ReviewEvent } from "../scientific-pedagogy/replay"; | ||
| import { fsrsIntervalDays, fsrsMemoryFromHistory, type ReviewEvent } from "../scientific-pedagogy/replay"; |
There was a problem hiding this comment.
The new runtime import of fsrsIntervalDays and fsrsMemoryFromHistory reaches directly into scientific-pedagogy/replay.ts, even though that module's explicit index.ts does not expose the replay implementation. This couples learning-studio to another module's internal evaluation file; publish a stable scheduler API or move the shared scheduling primitives behind an appropriate boundary.
AGENTS.md reference: AGENTS.md:L26-L26
Useful? React with 👍 / 👎.
| const reviewLog = [...(state.reviewLog ?? []), { cardId, at: now.toISOString(), outcome }]; | ||
| let next: ReviewState; | ||
| if (state.scheduler === "FSRS_4_5") { | ||
| const history = studioReviewEvents({ ...state, reviewLog }).filter((event) => event.itemId === cardId); |
There was a problem hiding this comment.
When a learner completes a quiz and then rates the same flashcard within 12 hours, studioReviewEvents puts the quiz result immediately before the new rating, and fsrsMemoryFromHistory ignores the rating as a same-session repeat. Consequently, a correct quiz followed by an explicit FAIL schedules the card from the quiz success (four days for a new card) instead of the failure (one day); rebuild scheduling history from flashcard ratings only, or otherwise ensure the current rating is always applied.
Useful? React with 👍 / 👎.
| const reviewLog = [...(state.reviewLog ?? []), { cardId, at: now.toISOString(), outcome }]; | ||
| let next: ReviewState; | ||
| if (state.scheduler === "FSRS_4_5") { | ||
| const history = studioReviewEvents({ ...state, reviewLog }).filter((event) => event.itemId === cardId); |
There was a problem hiding this comment.
After the global review log exceeds MAX_REVIEW_LOG, writeStudioState persists only its last 2,000 entries; following a reload, this reconstruction therefore loses older per-card events and can treat a previously mature but infrequently reviewed card as new, resetting its next interval on the next rating. Persist per-card FSRS memory or compact history per card rather than making scheduling depend on a globally truncated log.
AGENTS.md reference: AGENTS.md:L30-L31
Useful? React with 👍 / 👎.
Summary
Applies the recommendation from #19 without changing the default: Plan tab → Leitner (current, default) or FSRS-4.5 (experimental). FSRS rebuilds each card memory from its own review log (no new state, reversible) and schedules at 90 % predicted recall (interval = stability, 1–365 days).
Test plan
npm test631/631 (5 new), typecheck,eslint ., production build🤖 Generated with Claude Code