fix(quickadapter): isolate holdout and refit causal window (#115)

fix(quickadapter): isolate holdout and refit causal window (#115)

* fix(quickadapter): isolate holdout and refit causal window

* fix(quickadapter): align holdout selection and reporting

* fix(quickadapter): harden holdout/refit robustness and dedup imports (review fixes F1-F5)

* fix(quickadapter): migrate hp_rmse key and recompose inner-split weights (RC1,RC3)

* chore(quickadapter): drop hp_rmse config migration (assumed breaking change)

* docs(quickadapter): drop duplicated shuffle constraint from test_size row

* docs(quickadapter): correct test_size default to 0.1 (freqtrade-aligned, not template value)

* refactor(quickadapter): consolidate holdout_rmse init to a single site in fit()

* fix(quickadapter): guard infeasible inner validation split

Raise a contextual DependencyException when an integer test_size is not
smaller than the training rows left after the outer holdout, and when the
feature pipeline (SVM/DBSCAN outlier removal) empties the validation set
after transform, instead of surfacing a low-level sklearn/predict error.

* refactor(quickadapter): normalize None test_size in _add_refit_data

Match _add_validation_split: coerce a None test_size to _TEST_SIZE before
the ==0 short-circuit, so the refit gate stays consistent with the
validation gate regardless of the _TEST_SIZE value.