from sklearn.preprocessing import StandardScaler
from sklearn.svm import SVC

from sklearn.pipeline import Pipeline, make_pipeline



class FactoryEstimator(object):
    def __repr__(self):
        return "ConstructedEstimator(" + str(self.params) + ")"
    
    # for "is_classifier" ugh
    @property
    def steps(self):
        return self.constructor(self.params).steps
    
    def __init__(self, **params):
        self.params = params
        self.constructor = params['constructor']
    
    def get_params(self, deep=None):
        return self.params
    
    def set_params(self, **params):
        self.params.update(params)
        return self
        
    def fit(self, X, y):
        self.estimator_ = self.constructor(self.params)
        self.estimator_.fit(X, y)
        return self
    
    def predict(self, X):
        return self.estimator_.predict(X)
    

func = lambda x: make_pipeline(StandardScaler(with_mean=x['with_mean']), SVC(gamma=x['gamma'], C=x['C']))
est = FactoryEstimator(constructor=func, with_mean=True, C=1, gamma=.1)

grid = GridSearchCV(est, param_grid={'C': [.1, 1, 10], 'gamma': [0.001, 0.1, 1, 0]}, scoring='accuracy')
grid.fit(X, y)
print(grid.best_score_)

est = ConstructedEstimator(constructor=func, with_mean=True, C=1, gamma=.1)

from sklearn.datasets import load_iris

iris = load_iris()

X, y = iris.data, iris.target

est.fit(X, y)

est.predict(X)

from sklearn.grid_search import GridSearchCV



grid.fit(X, y)

grid.best_estimator_

grid.best_estimator_

grid.predict(X)

from sklearn.cross_validation import cross_val_score

est = ConstructedEstimator(constructor=func, with_mean=True, C=1, gamma=.1)

cross_val_score(est, X, y, scoring="accuracy")

est.fit(X, y)

bla = func({'C':1, 'with_mean': True, 'gamma': 0.1})

bla.fit(X, y)

bla.score(X, y)

cross_val_score(bla, X, y, scoring="accuracy")

est.steps_