- Remove hardcoded database_url/jwt_secret defaults, require env vars
- Add DB indexes for stocks.market, market_cap, backtests.user_id
- Optimize backtest engine: preload all prices, move stock_names out of loop
- Fix backtest API auth: filter by user_id at query level (6 endpoints)
- Add manual transaction entry modal on portfolio detail page
- Replace console.error with toast.error in signals, backtest, data explorer
- Add backtest delete button with confirmation dialog
- Replace simulated sine chart with real snapshot data
- Add strategy-to-portfolio apply flow with dialog
- Add DC pension risk asset ratio >70% warning on rebalance page
- Add backtest comparison page with metrics table and overlay chart
- Add WalkForwardResult model with train/test window metrics
- Create WalkForwardEngine that reuses existing BacktestEngine
with rolling train/test window splits
- Add POST/GET /api/backtest/{id}/walkforward endpoints
- Add Walk-forward tab to backtest detail page with parameter
controls, cumulative return chart, and window results table
- Add Alembic migration for walkforward_results table
- Add 8 unit tests for window generation logic (100 total passed)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add min_trade_amount parameter (default 10,000 KRW) to rebalance/calculate
endpoint. Trades below this threshold are converted to hold actions to avoid
inefficient micro-trades during rebalancing.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Validates trading day count, benchmark coverage, per-date ticker
density, and date gaps before running simulation. Logs warnings for
holdings with missing prices during execution.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Port make-quant-py's FnGuide scraping logic into galaxy-po's
BaseCollector pattern. Collects annual and quarterly financial
statements (revenue, net income, total assets, etc.) and maps
Korean account names to English keys for FactorCalculator.
Scheduled weekly on Monday 19:00 KST since data updates quarterly.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
pykrx get_market_ohlcv returns 6 data columns (시가/고가/저가/종가/거래량/거래대금),
not 7. The 등락률 (change) column does not exist, causing a length mismatch error.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The pykrx library uses 'etx' not 'etf' as the module directory name,
and fetch() is the proper method that returns a DataFrame.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
KRX CSV download endpoint blocks requests from cloud/server IPs,
causing "No columns to parse from file" errors. Replaced with pykrx's
JSON-based API (get_market_ticker_list, get_market_cap_by_ticker,
get_market_fundamental_by_ticker) which is more reliable.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add ETFCollector (KRX master data) and ETFPriceCollector (pykrx OHLCV)
with corresponding admin API endpoints and frontend collection UI buttons.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- APScheduler for daily snapshots (18:30 weekdays)
- ReturnsCalculator with CAGR, TWR, MDD, volatility
- Portfolio history page with snapshots and returns tabs
- FastAPI lifespan integration for scheduler
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- VirtualPortfolio for portfolio simulation
- BacktestEngine for strategy backtesting
- Worker for async background execution
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- BaseStrategy abstract class
- MultiFactorStrategy with weighted factors
- QualityStrategy with F-Score filtering
- ValueMomentumStrategy combining value and momentum
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add time import and RATE_LIMIT_DELAY constant for rate limiting between HTTP requests
- Add 1-second delay after OTP request to respect API rate limits
- Validate OTP response is not empty before using it
- Add CSV column structure validation with required columns check
- Add data quality check to skip records where all metrics are None
- Improve error handling and data integrity
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add ValuationCollector class that fetches PER, PBR, and dividend yield
data from KRX for all listed stocks. Includes business day validation,
safe float conversion, and upsert logic for the valuations table.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Implement per-ticker commits to ensure atomic operations per data source
- Add rollback on exception to prevent partial data corruption
- Add _safe_float() and _safe_int() helper methods for defensive type conversion
- Validate column count after DataFrame reset to catch schema issues early
- Skip records with missing essential values (close price) with debug logging
- Remove final db.commit() since commits now happen per ticker in the loop
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add REQUEST_TIMEOUT and RATE_LIMIT_DELAY constants to StockCollector
- Add timeout parameter to all HTTP requests
- Wrap HTTP requests in try-except with proper error handling
- Add _validate_biz_day() method to both collectors
- Add validation for required fields (ticker, name) before insert
- Replace generic Exception with specific exception types in SectorCollector
- Add logging module and logger to both collectors
- Remove unused numpy import from StockCollector
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Implement StockCollector to fetch stock master data from KRX
(Korea Exchange) including market cap, EPS, BPS, and dividend info.
Implement SectorCollector to fetch WICS sector classification from
WISEindex. Both collectors use PostgreSQL upsert for efficient updates.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add session validation in __init__ to ensure database session is not None
- Implement transaction rollback handling in complete_job() for exception safety
- Implement transaction rollback handling in fail_job() for exception safety
- Improve exception handling in run() to gracefully handle fail_job failures while preserving original exception
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
