Add linear regression stats of log-equity curve

[adrianhasse]

No description provided.

[Copilot]

Pull Request Overview

This PR adds linear regression analysis of log-equity curves to compute statistical metrics for portfolio performance evaluation. The changes include calculating slope, standard error, t-statistic, and R² values for both portfolio and benchmark equity curves.

Key changes:

• Adds new _equity_curve_stats method to perform linear regression on log-equity curves
• Updates benchmark metric naming for consistency (total_benchmark_return → benchmark_total_return)
• Adds scipy dependency for statistical calculations

Reviewed Changes

Copilot reviewed 4 out of 5 changed files in this pull request and generated 1 comment.

File	Description
kissbt/analyzer.py	Implements the new _equity_curve_stats method and integrates it into performance metrics calculation
pyproject.toml	Adds scipy dependency for statistical calculations
tests/test_analyzer.py	Adds comprehensive tests for the new linear regression statistics functionality
tests/test_integration.py	Updates test assertions to use the new benchmark metric naming

Comments suppressed due to low confidence (2)

kissbt/analyzer.py:100

• [nitpick] The error message is clear but could be more specific about which series contains the non-positive values. Consider adding context like 'Portfolio value series contains...' or 'Benchmark value series contains...' to help with debugging.

            raise ValueError(

tests/test_analyzer.py:36

• [nitpick] The magic number calculation 256 * 3 is unclear. Consider using a more descriptive variable name or comment to explain why 768 days (approximately 3 trading years) is used.

    num_days = 256 * 3