| Age | Commit message (Collapse) | Author |
|---|
|
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Reduce layout shift during data loading by using content-shaped
skeleton placeholders instead of generic spinners in DeckDetailPage,
DeckCardsPage, and StudyPage.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Implement study day boundary at 3:00 AM for due card logic
|
|
Instead of comparing due timestamps exactly (card.due <= now), compare
against the next 3 AM boundary so all cards due within the current
study day appear at once. This prevents new cards from trickling in
throughout the day when FSRS fuzz spreads due times.
https://claude.ai/code/session_01FeDztLcyGofd6nxh8ct7a3
|
|
|
|
|
|
|
|
Prevent cards with similar intervals from always appearing on the same
day by enabling ts-fsrs built-in fuzz factor.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Allow continuous study flow by pressing Space repeatedly - first to
flip the card, then to rate as Good (3).
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Replace custom createReloadableAtom/createReloadableAtomFamily with
atomWithSuspenseQuery from jotai-tanstack-query, leveraging TanStack
Query's built-in caching, invalidation, and Suspense support.
- Add @tanstack/query-core and jotai-tanstack-query dependencies
- Create shared QueryClient instance (src/client/queryClient.ts)
- Migrate all data atoms (decks, cards, study, noteTypes) to atomWithSuspenseQuery
- Update page components to use .data destructuring and queryClient.invalidateQueries()
- Update all test files to use QueryClient for data hydration
- Remove src/client/atoms/utils.ts (no longer needed)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Add a debounced (500ms) search input using use-debounce that filters
note groups by matching card front/back content (case-insensitive).
Pagination resets to page 1 when the search query changes.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
|
|
Separate deck detail and card learning pages
|
|
Separate the card list view from the deck learning page to prevent users
from seeing cards they are about to study. The deck detail page now shows
only study statistics with a "Study Now" button and a "View Cards" link.
- Add new DeckCardsPage component at /decks/:deckId/cards for managing cards
- Simplify DeckDetailPage to show deck stats and navigation buttons
- Update routing in App.tsx with proper route ordering
- Add comprehensive tests for both pages
|
|
Display daily card count on deck list page
|
|
|
|
|
|
Display a badge with the number of cards due for study today next to
each deck name on the home page. The count is fetched along with deck
data from the API to minimize additional network requests.
|
|
Replace AuthProvider and SyncProvider with Jotai atoms for more granular
state management and better performance. This migration:
- Creates atoms for auth, sync, decks, cards, noteTypes, and study state
- Uses atomFamily for parameterized state (e.g., cards by deckId)
- Introduces StoreInitializer component for subscription initialization
- Updates all components and pages to use useAtomValue/useSetAtom
- Updates all tests to use Jotai Provider with createStore pattern
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
|
|
Cards are now randomized using Fisher-Yates algorithm to improve
learning by preventing users from memorizing card order.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Show actual note types and their fields in the expected format section
instead of a hardcoded example. Fields are sorted by order.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Since import is initiated from deck context, the deck is already known
via props. Simplifies CSV format to: note_type,Field1,Field2,...
BREAKING CHANGE: CSV format changed from deck,note_type,... to note_type,...
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Replace raw fetch() calls with apiClient.rpc typed client across all
modal and page components. This provides better type safety and
eliminates manual auth header handling.
- Make handleResponse public for component usage
- Update all component tests to mock RPC methods instead of fetch
- Change POSTGRES_HOST default to kioku-db for Docker compatibility
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Document the CSV import feature including format specification,
examples, and troubleshooting guide.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
|
|
Add client-side CSV parsing and bulk import API endpoint for importing
notes from CSV files. Supports quoted fields, newlines in values, and
escaped quotes.
- New POST /api/decks/{deckId}/notes/import endpoint for bulk creation
- CSV parser with RFC 4180 compliance
- Multi-phase import modal (upload → validate → preview → import)
- Client-side validation with per-row error reporting
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Change template input fields from single-line <input> to <textarea>
elements to support multiline content in front/back templates.
Fixes #5
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
When the refresh token fails (session expired), the ApiClient now
notifies the AuthProvider via a callback. This triggers a logout
and React state update, causing ProtectedRoute to redirect to /login.
Closes #7
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
The client session was too short because access tokens (15 min) weren't
being automatically refreshed using the refresh token (7 days). Now the
ApiClient intercepts 401 responses, attempts token refresh, and retries
the original request. This extends effective session duration to 7 days.
Closes #6
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
|
|
Ensures deterministic ordering for all multi-row SELECT queries:
- deck/note/noteType findByUserId/findByDeckId: order by createdAt
- card findByNoteId: order by isReversed (normal card first)
- note field values: order by noteFieldTypeId
- sync pull queries: order by id
This guarantees consistent UI display and sync results regardless
of PostgreSQL's internal row ordering.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
|
|
The empty state sections already have create buttons in the page header,
making these inline buttons redundant.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Replace manually defined AuthResponse and User types with Hono's
InferResponseType to automatically derive types from server definitions.
This ensures client types stay in sync with server responses.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Server returns `{ error: { message, code } }` but client expected
`{ error: string, code }`, causing "[object Object]" to display
on login failure.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Eliminates duplicate configuration by building the connection URL
from POSTGRES_USER, POSTGRES_PASSWORD, POSTGRES_DB, POSTGRES_HOST,
and POSTGRES_PORT instead of requiring a separate DATABASE_URL.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
|
|
Remove ConflictResolutionItem type since resolution field was always
"server_wins" and provided no useful information. Resolution arrays
now contain entity IDs directly instead of objects.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Remove the unused "local_wins" strategy and ConflictResolverOptions
interface from ConflictResolver. The CRDT-based conflict resolution
now always uses Automerge merge with server_wins fallback when CRDT
data is unavailable. This simplifies the codebase by removing
configuration options that were never used in production.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Remove the deprecated "newer_wins" strategy from the conflict resolver
as CRDT is now the default and validated approach. The system now uses
only "crdt", "server_wins", and "local_wins" strategies.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Automerge uses WebAssembly which requires vite-plugin-wasm and
vite-plugin-top-level-await to handle WASM modules in production builds.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
No server-side data migration script needed since there is no existing
production data. The CRDT schema is already in place from Phase 4.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Add one-time migration script to convert existing local IndexedDB
entities to Automerge CRDT documents. Includes:
- Migration with idempotency check (runs only once)
- Batch processing option for large datasets
- Progress callback for UI feedback
- Unit tests for migration logic
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Add comprehensive test suite covering multi-device concurrent editing:
- Two devices editing different fields (no conflicts)
- Two devices editing same field (deterministic resolution)
- Concurrent edit and delete scenarios
- Card FSRS field concurrent updates
- Incremental sync simulation
- Three-way merge scenarios
- Serialization roundtrip with concurrent edits
- Edge cases (null handling, rapid edits)
- Offline queue simulation
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Add crdtChanges field to sync push/pull endpoints for CRDT document
synchronization. The server now stores and retrieves CRDT binaries
from the crdt_documents table, enabling conflict-free sync between
clients.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Add Drizzle migration to create the crdt_documents table with:
- UUID primary key with auto-generation
- Foreign key to users table with cascade delete
- Unique index on (user_id, entity_type, entity_id)
- Indexes for entity_type and sync_version queries
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
Add PostgreSQL schema for storing Automerge CRDT documents with
indexes for efficient querying by user, entity type, and sync version.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
|
- Store CRDT document binaries after successful push operations
- Update CRDT sync metadata (lastSyncAt, syncVersionWatermark) after sync
- Add getCrdtSyncStats(), clearCrdtState(), hasCrdtDocument() methods
- Add crdt_documents_stored event and crdtDocumentsStored to SyncResult
- Include all entity types in conflict resolution count
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
