feat(strategy-v2): expand FactorResearchEngine with multi-family alpha factors

[huuan26]

Summary

Expands FactorResearchEngine with vectorized cross-sectional panel implementations across major Alpha factor families:

• Statistical moments (Rolling Skewness, Rolling Kurtosis)
• Liquidity & Volume dynamics (Amihud Illiquidity, Volume Ratio, Turnover Proxy, CMF, MFI, Efficiency Ratio, OBV Slope)
• Volatility & Tail risk (Normalized ATR, Ulcer Index, Downside Volatility, Bollinger Bandwidth, Max Drawdown)
• Momentum & Oscillators (Wilder RSI, CMO, MACD Histogram, SMA/EMA Distance)
• Expanded fundamental valuation & quality mappings (book_to_price, earnings_yield, debt_to_equity, revenue_growth, free_cash_flow)

Related issue

• Related to [Roadmap] Dynamic universe and point-in-time stock screening #251, [Roadmap] Portfolio research and execution roadmap #257 (Portfolio Research and Execution milestone)

Changes

• Vectorized 2D panel calculations across the asset cross-section in FactorResearchEngine._factor_panel().
• Expanded _FUNDAMENTAL_FIELDS and required_fields() resolution for fundamental multi-factor screening.
• Enriched _CORRELATION_FACTORS to represent liquidity, distribution moments, and momentum style dimensions.
• Guarded _factor_correlation() against zero-variance rank correlation edge cases.
• Added comprehensive unit tests in tests/test_factor_research.py covering factor diagnostics, group curve formation, and point-in-time lookahead protection.

Test plan

• Relevant pytest tests pass (pytest tests/test_factor_research.py tests/test_factor_registry.py - 23 passed)
• git diff --check passed with 0 errors
• Invariant tests verify zero lookahead leakage across historical snapshots

API documentation (if routes/schemas changed)

• No public HTTP routes or OpenAPI schemas modified. All additions are additive within FactorResearchEngine.