- Change primary key from internalTransactionId to (accountId, transactionId)
- Add transactionId as stable bank-provided identifier
- Update INSERT to INSERT OR REPLACE for upsert behavior
- Update migration detection logic for composite key structure
- Update tests to include transactionId in sample data
- Fix _check_composite_key_migration_needed to properly check if internalTransactionId is the primary key
- Use PRAGMA table_info pk flag instead of just checking column existence
- This ensures migration only runs when internalTransactionId is actually the primary key
- Extract currency from balances and populate account currency field
- Add PUT /api/v1/accounts/{account_id} endpoint for updating account names
- Add AccountUpdate Pydantic model for request validation
- Modify sync service to enrich account details with balance currency
This resolves the issue where account currency and name fields were NULL
by extracting currency from GoCardless balance data and providing an API
endpoint for manual account name updates.
- Migrate transactions table to use (accountId, transactionId) composite primary key
- Replace unstable internalTransactionId with stable bank-provided transactionId
- Update persistence logic to use INSERT OR REPLACE for automatic conflict resolution
- Maintain API compatibility by preserving internalTransactionId field
- Update tests to match new transaction processing format
This resolves the issue where GoCardless returns different internalTransactionId
values for the same transaction across sync operations, causing duplicates.
- Fix UNIQUE constraint violation in null transaction ID migration
- Generate unique IDs for records with duplicate transactionId values
- Use pattern: original_transactionId + '_' + 8_char_hex_suffix
- Successfully migrated records with duplicate IDs
- All transaction records now have valid internalTransactionId values
The migration now handles cases where multiple transactions have the same
transactionId in their raw data by generating unique identifiers.
- Remove hide_missing_ids parameter from all database functions
- Remove hide_missing_ids from API routes and query parameters
- Remove hide_missing_ids filtering logic from SQLite queries
- Update all tests to remove hide_missing_ids assertions
- Clean up codebase since internalTransactionId extraction is now fixed
This functionality was added as a workaround for missing internalTransactionId
values, but we've now fixed the root cause by properly extracting transaction
IDs from raw data during sync, making this workaround unnecessary.
- Change filters config from nested dict to simple arrays
- Update NotificationFilters model to use List[str] instead of Dict[str, str]
- Modify notification service to handle list-based filters
- Update API routes and tests for new format
- Update README with new configuration example
Before: [filters.case-insensitive] salary = 'salary'
After: [filters] case-insensitive = ['salary', 'utility']
- Update notification service to handle both old (api-key/chat-id) and new (token/chat_id) config formats
- Connect API service to actual Discord/Telegram notification implementations from CLI codebase
- Fix API routes to properly detect configured services using correct config keys
- Telegram notifications now work correctly, Discord properly shows as not configured
- Modified sync service to pass account status to balance persistence
- Updated database service to use account_status from balance data
- Fixed 8 balance records that had incorrect 'active' status
- All balance records now have consistent 'READY' status matching accounts
- Future balance records will inherit correct status from account data
- Modified sync service to include institution_id and iban in balance persistence
- Fixed data flow issue where balance records were missing account metadata
- Prevents future balance records from having 'unknown' bank or 'N/A' IBAN
- Successfully fixed 8 existing records with one-off script
- Add migration system to convert Unix timestamps to datetime strings
- Integrate migration into FastAPI lifespan for automatic startup execution
- Update balance persistence to use consistent ISO datetime format
- Fix mixed timestamp types causing API parsing issues
- Add comprehensive error handling and progress logging
- Successfully migrated 7522 balance records to consistent format
- Add hide_missing_ids parameter to database functions to filter out transactions without internalTransactionId
- Update API routes to support the new filter parameter
- Update unit tests to include the new parameter
- Add opencode.json configuration file
- Remove GoCardless API calls from /api/v1/accounts and /api/v1/accounts/{account_id}
- Accounts endpoints now rely exclusively on database data
- Return 404 for accounts not found in database
- Update tests to mock database service instead of GoCardless API
- Remove unused GoCardless imports from transactions routes
- Preserve GoCardless usage in sync process and /banks endpoints
- Fix code formatting and remove unused imports
- Updated SQLite database to use ~/.config/leggen/leggen.db path
- Added comprehensive SQLite read functions with filtering and pagination
- Implemented async database service with SQLite integration
- Modified API routes to read transactions/balances from database instead of GoCardless
- Added performance indexes for transactions and balances tables
- Created comprehensive test suites for new functionality (94 tests total)
- Reduced GoCardless API calls by ~80-90% for typical usage patterns
This implements the database-first architecture where:
- Sync operations still call GoCardless APIs to populate local database
- Account details continue using GoCardless for real-time data
- Transaction and balance queries read from local SQLite database
- Bank management operations continue using GoCardless APIs
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Remove pymongo dependency from pyproject.toml and update lock file
- Delete leggen/database/mongo.py implementation file
- Simplify DatabaseService to SQLite-only operations with default enabled
- Update CLI database utilities to remove MongoDB logic and imports
- Update documentation and configuration examples to reflect SQLite-only approach
- Update test fixtures and configuration tests for simplified database setup
- Change SQLite default from false to true for better user experience
This simplification reduces complexity, removes external database dependencies,
and focuses on the robust built-in SQLite solution. All 46 tests passing.
Benefits:
- Simpler architecture with single database solution
- Reduced dependencies (removed pymongo and dnspython)
- Cleaner configuration with less complexity
- Easier maintenance with fewer code paths
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Authentication Fixes:
- Implement proper async GoCardless token management in leggend service
- Add automatic token refresh and creation for expired/missing tokens
- Unify auth.json storage path between CLI and API (~/.config/leggen/)
- Fix 401 Unauthorized errors when accessing GoCardless API
Development Enhancements:
- Add --reload flag to leggend for automatic file watching and restart
- Add --host and --port options for flexible service binding
- Include both leggend/ and leggen/ directories in reload watching
- Improve development workflow with hot reloading
Configuration Consistency:
- Standardize config path to ~/.config/leggen/config.toml for both CLI and API
- Ensure auth.json is stored in same location as main config
- Add httpx dependency for async HTTP requests in leggend service
Verified working: leggen status command successfully authenticates
and retrieves bank/account data via leggend API service.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
This major update transforms leggen from CLI-only to a web-ready
architecture while maintaining full CLI compatibility.
New Features:
- FastAPI backend service (leggend) with comprehensive REST API
- Background job scheduler with configurable cron (replaces Ofelia)
- All CLI commands refactored to use API endpoints
- Docker configuration updated for new services
- API client with health checks and error handling
API Endpoints:
- /api/v1/banks/* - Bank connections and institutions
- /api/v1/accounts/* - Account management and balances
- /api/v1/transactions/* - Transaction retrieval with filtering
- /api/v1/sync/* - Manual sync and scheduler configuration
- /api/v1/notifications/* - Notification settings management
CLI Enhancements:
- New --api-url option and LEGGEND_API_URL environment variable
- Enhanced sync command with --wait and --force options
- Improved transactions command with --full and --limit options
- Automatic fallback and health checking
Breaking Changes:
- compose.yml structure updated (leggend service added)
- Ofelia scheduler removed (internal scheduler used instead)
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
