• from sklearn.pipeline import Pipeline
    from sklearn.base import TransformerMixin
    import numpy as np
    state = 0
    class Cacher(TransformerMixin):
        def __init__(self, internal=None, cache=None):
            self.internal = internal
            self.cache = cache or dict()
            assert internal
        def fit(self, data, y=None):  
            """Returns `self` unless something different happens in train and test"""
            return self
        def save(self):
            '''Save to disk...'''
            return True
        def transform(self, X, y=None):
            def cacheado(i):
                hash = str(i)
                if hash not in self.cache:
                    self.cache[hash] = self.internal.transform(i)
                return self.cache[hash]
            res = np.apply_along_axis(cacheado, 0, X)
            return res
    class Transformer(TransformerMixin):    
        def fit(self, data, y=None):
            return self
        def transform(self, X, y=None):
            print('Original transformer')
            return X+state
    # DO IT ONCE
    ner_pipe = Pipeline([
        ('trans', Cacher(internal=Transformer()))
    for i in range(10):
        state = i
        print(ner_pipe.transform(np.array([[0,1], [1, 2]])))
        if i % 4 == 0:
            ner_pipe.fit([[0,1], [1, 2]])
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment