How to use the contextualbandits.utils._BootstrappedClassifierBase function in contextualbandits

def exploit(self, X):
        return self._score_avg(X)

    def predict(self, X):
        ### Thompson sampling
        if self.percentile is None:
            pred = self._score_rnd(X)

        ### Upper confidence bound
        else:
            pred = self._score_max(X)

        return pred

class _BootstrappedClassifier_w_predict_proba(_BootstrappedClassifierBase):
    def _get_score(self, sample, X):
        return self.bs_algos[sample].predict_proba(X)[:, 1]

class _BootstrappedClassifier_w_decision_function(_BootstrappedClassifierBase):
    def _get_score(self, sample, X):
        pred = self.bs_algos[sample].decision_function(X).reshape(-1)
        _apply_sigmoid(pred)
        return pred

class _BootstrappedClassifier_w_predict(_BootstrappedClassifierBase):
    def _get_score(self, sample, X):
        return self.bs_algos[sample].predict(X).reshape(-1)

class _OneVsRest:
    def __init__(self, base, X, a, r, n, thr, alpha, beta, smooth=False, assume_un=False,
                 partialfit=False, force_fit=False, force_counters=False, njobs=1):

else:
            pred = self._score_max(X)

        return pred

class _BootstrappedClassifier_w_predict_proba(_BootstrappedClassifierBase):
    def _get_score(self, sample, X):
        return self.bs_algos[sample].predict_proba(X)[:, 1]

class _BootstrappedClassifier_w_decision_function(_BootstrappedClassifierBase):
    def _get_score(self, sample, X):
        pred = self.bs_algos[sample].decision_function(X).reshape(-1)
        _apply_sigmoid(pred)
        return pred

class _BootstrappedClassifier_w_predict(_BootstrappedClassifierBase):
    def _get_score(self, sample, X):
        return self.bs_algos[sample].predict(X).reshape(-1)

class _OneVsRest:
    def __init__(self, base, X, a, r, n, thr, alpha, beta, smooth=False, assume_un=False,
                 partialfit=False, force_fit=False, force_counters=False, njobs=1):
        if 'predict_proba' not in dir(base):
            base = _convert_decision_function_w_sigmoid(base)
        if partialfit:
            base = _add_method_predict_robust(base)
        if isinstance(base, list):
            self.base = None
            self.algos = base
        else:
            self.base = base
            self.algos = [deepcopy(base) for i in range(n)]

def predict(self, X):
        ### Thompson sampling
        if self.percentile is None:
            pred = self._score_rnd(X)

        ### Upper confidence bound
        else:
            pred = self._score_max(X)

        return pred

class _BootstrappedClassifier_w_predict_proba(_BootstrappedClassifierBase):
    def _get_score(self, sample, X):
        return self.bs_algos[sample].predict_proba(X)[:, 1]

class _BootstrappedClassifier_w_decision_function(_BootstrappedClassifierBase):
    def _get_score(self, sample, X):
        pred = self.bs_algos[sample].decision_function(X).reshape(-1)
        _apply_sigmoid(pred)
        return pred

class _BootstrappedClassifier_w_predict(_BootstrappedClassifierBase):
    def _get_score(self, sample, X):
        return self.bs_algos[sample].predict(X).reshape(-1)

class _OneVsRest:
    def __init__(self, base, X, a, r, n, thr, alpha, beta, smooth=False, assume_un=False,
                 partialfit=False, force_fit=False, force_counters=False, njobs=1):
        if 'predict_proba' not in dir(base):
            base = _convert_decision_function_w_sigmoid(base)
        if partialfit:
            base = _add_method_predict_robust(base)