External Sources:
- New ExternalSources.py module for real-time API data fetching
- Configurable refresh intervals, JSONPath extraction, and auth headers
- External sources can be used in signals for live strategy conditions
External Indicators:
- New ExternalIndicators.py module for historical API data
- Supports backtesting by fetching and caching historical data points
- Timestamps are matched to candle times during backtest execution
- Fear & Greed index and similar APIs now work in backtests
Signal System Enhancements:
- Added signal Blockly blocks for use in strategy builder
- New signal_blocks.js for dynamic signal block generation
- Signals category added to strategy toolbox
- PythonGenerator updated to handle signal_* block types
- process_signal() method added to StrategyInstance classes
notify_user Block Enhancement:
- Now supports dynamic values (variables, indicators, expressions)
- Can display indicator values and calculations in notifications
Bug Fixes:
- Fixed Socket.IO parse error causing strategies not to load on refresh
- Enhanced sanitize_for_json to handle pandas, datetime, bytes types
- Fixed external indicator name matching (underscore/space conversion)
- Fixed PRAGMA query missing fetch_all=True in ExternalIndicators
Dependencies:
- Added jsonpath-ng for JSONPath extraction from API responses
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Implement a subscription system for public strategies:
Security & Auth:
- Bind WebSocket user identity at connect time (prevents spoofing)
- Add server-side ownership verification for all destructive operations
- Invalidate socket connections on logout
- Add XSS protection with HTML/JS escaping in frontend
Database:
- Add strategy_subscriptions table with proper indexes
- Fix get_all_rows_from_datacache to fall back to DB when cache empty
Backend:
- Add subscribe/unsubscribe endpoints with authorization checks
- Add get_user_strategies (owned + subscribed) and get_public_strategies_catalog
- Propagate indicator_owner_id through strategy instances for subscribed strategies
- Redact strategy internals (code, workspace) for non-owners
Frontend:
- Add "Add Public" button to browse and subscribe to public strategies
- Show subscribed strategies with creator badge and unsubscribe button
- Prevent editing of subscribed strategies (show info modal instead)
- Add public strategy browser modal
Tests:
- Update authorization tests for new subscription-required model
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add UTC time formatter to charts for consistent time display
- Show both local and UTC times in backtest trades table
- Add indicator warmup period calculation to fetch extra candles
- Append 'Z' suffix to trade/alert timestamps to indicate UTC
- Add get_available_balance to strategy execution context
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add Open/Close Time columns to Trades Executed table
- Display trade markers (buy/sell arrows) automatically on chart
- Clear markers when closing results dialog or running new test
- Collect notify_user alerts with timestamps during backtesting
- Display Strategy Alerts section in backtest results
- Fix timestamp conversion: use unit='s' for EDM timestamps (not 'ms')
- Fix trade datetime extraction using data feed's datetime method
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
EDM Integration:
- Add edm_client module for REST/WebSocket communication with exchange-data-manager
- Refactor candles.py to fetch data exclusively from EDM
- Add EDM session management for authenticated exchange access
- Update backtesting to use EDM for historical data
- Pass edm_client through to all strategy instances
- Add /api/edm-config, /api/chart-view, /health/edm endpoints
- Remove legacy candle-fetching code from DataCache_v3 and Exchange
UI Improvements:
- Make all popup dialogs draggable and resizable (exchange config, indicator, signal, trade)
- Add default trading source (exchange, symbol, timeframe) to strategy dialog
- Convert symbol fields to dropdowns with smart prioritization (popular pairs first)
- Fix signal dialog property dropdown initialization
Bug Fixes:
- Fix PythonGenerator regex to handle symbols starting with numbers (e.g., 0G/USDT)
- Fix EDM client date format (use Unix timestamps instead of ISO strings)
- Fix strategy default_source database migration
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Create brokers package with unified trading interface
- BaseBroker: Abstract base class defining broker contract
- BacktestBroker: Wraps Backtrader for backtesting mode
- PaperBroker: Simulated order execution for paper trading
- Factory function to create broker based on trading mode
- Comprehensive test suite for broker functionality
The broker abstraction enables strategies to work identically
across backtest, paper, and live trading modes.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
rob