Feature/20260708 product m u 2 - #48
Merged
Merged
Conversation
Replace hardcoded [1,3,7,14,30] day intervals with the SM-2 spaced repetition algorithm. EF and interval are now persisted per thought and adjusted by answer quality (Passed/Sloppy/Failed). - Add SrsState, ChallengeQuality, sm2_next() in thought_parser - DB migration v2→v3: srs_easiness_factor, srs_interval_days columns - Rewrite next_due_after_anchor() to use SM-2 interval with legacy fallback - Sort review queue by EF ascending (harder first) - Frontend: call write-back on all outcomes, pass sloppy flag - 6 new SM-2 unit tests
Also fix settings modal to allow clearing API key. Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Pure-backend latent paragraph marking with three strategies: high-similarity pairs, semantic isolation, cross-doc recurrence. Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
… queue Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
…ry export - Add cognitive push scheduler (startup + 30min periodic) with tauri-plugin-notification - Add configurable push frequency and quiet hours in Settings > Cognition - Add ThoughtGrowthStoryCard component with journey timeline and Markdown export - Add get_thought_growth_story Tauri command for reading thought history - Register growth_story module and cognitive_push in lib.rs invoke_handler Co-authored-by: Claude (big-pickle) <noreply@anthropic.com>
…hancements Add AI-not-configured guide to writing coach and thought AI panel; add image export, privacy checks and confirm dialog for growth story. Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
…empty Previously scan_vault only ran after rebuild_index with indexed_chunks > 0. If the embedding index was already built, subsequent app launches never triggered a scan, leaving the review queue permanently empty of latent paragraph candidates. Now open_workspace checks for the "has chunks but no candidates" condition and fires a background scan to populate the candidates table. Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Review panel: - Fix React key collision for candidate items (candidateId || thoughtId) - Add loading state text on "Start challenge" button while generating - Add "Skip" button that dismisses current candidate and reloads queue - Strip .md extension from candidate file display name - Show marking reason (similar/isolated/cross-doc) below candidate label Content filters (latent_paragraphs.rs): - Add table filter: skip chunks where >50% lines contain pipe chars - Add frontmatter filter: skip YAML metadata blocks (--- fenced) - Add heading-only filter: skip chunks with only # heading lines - Add 6 unit tests for the new filters Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
…code) Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
For cross_doc_recurrence candidates, store the other document paths in the cluster (excluding self) so the UI can display which notes contain similar concepts. For high_similarity candidates, the paired document was already available. Frontend shows a "Related notes:" section with document name tags below the marking reason when viewing a candidate in the review panel. Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Add a prefetch pipeline that generates questions in the background: - On queue load, immediately start generating for items[0], chain to [1] - startRound checks cache first — cache hit = instant question (0 wait) - Cache miss falls back to on-demand generation (existing behavior) - After answering or skipping, prefetch the next item in the queue - Cache is cleared on queue reload (dismiss/promote) to stay fresh Typical user experience after this change: - First item: 0-5s wait (prefetch races with user reading the excerpt) - Second item onward: 0 wait (question pre-generated while answering) Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Store in-flight prefetch Promises in a Map. When the user clicks "Start Challenge" while a prefetch is still pending, await the existing Promise instead of firing a duplicate LLM request. Three-tier lookup in startRound: 1. Cache hit → instant (0 wait, 0 requests) 2. In-flight hit → await same Promise (reduced wait, 0 extra requests) 3. Miss → on-demand generation (full wait, 1 request) Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Users can now exit a challenge question without answering by clicking "Give up" / "放弃本题", which resets to the pick phase. Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Problem: candidates generated before new content filters (table, frontmatter, heading-only) were added remain in the DB and still show table-heavy content in the review queue. Solution: - Add FILTER_VERSION constant (bumped to 2) and latent_meta table - On workspace open, check stored version vs current; if mismatched, clear non-dismissed candidates and re-scan with updated filters - Strengthen table detection: any line with |---| triggers skip, or >30% pipe-containing lines (was 50%) Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
摘要
变更类型
检查清单
npm run build（若改动前端）cd src-tauri && cargo test（若改动 Rust）备注