Daily Audit 2026-04-24: Fix CatBoost optimizer bug, print() violation, hardcoded capital

[Copilot]

Description

Three new bugs identified during the 2026-04-24 codebase audit: a logic error that feeds LightGBM hyperparameters into CatBoost, a print() call violating the Logger singleton convention, and a hardcoded initial capital value that diverges from the actual simulation config.

Changes in this pull request

• src/pipeline/single_pipeline.py — CatBoost gets wrong hyperparameters when optimisation is enabled
  The original else branch ran LightGBM Optuna optimisation for any non-XGBoost model type, passing LightGBM params (num_leaves, bagging_fraction, …) into CatBoostStockPredictor which expects depth/iterations/l2_leaf_reg. Now uses explicit elif "lightgbm" branch; "catboost" (and any future type) logs a warning and falls back to model defaults. optimizer.save_results() deduplicated to one call.

  # before
  if self.model_type == "xgboost":
      optimizer.optimize_xgboost()
      self._model_params = optimizer.best_xgb_params
  else:                              # catboost silently lands here
      optimizer.optimize_lightgbm()
      self._model_params = optimizer.best_lgb_params  # wrong params for CatBoost
  
  # after
  if self.model_type == "xgboost":
      optimizer.optimize_xgboost()
      self._model_params = optimizer.best_xgb_params
  elif self.model_type == "lightgbm":
      optimizer.optimize_lightgbm()
      self._model_params = optimizer.best_lgb_params
  else:
      logger.warning("No Optuna optimizer available for '%s'. Using defaults.", self.model_type)
      self._model_params = None
      return
  optimizer.save_results(...)

• src/ranking/ranker.py — print() instead of logger.warning()
print("⚠️ No stocks meet the confidence threshold…") replaced with logger.warning(…); added missing from src.util import logger import.

• src/writer/result_writer.py — hardcoded $10,000.00 ignores actual simulation capital
  persist_results now accepts initial_capital: float = 10_000.0 (default preserves backward compatibility) and writes f" Initial Capital: ${initial_capital:,.2f}\n\n".