Coverage for mlprodict/grammar/grammar_sklearn/g_sklearn_preprocessing.py: 100%

1# -*- coding: utf-8 -*- 

2""" 

3@file 

4@brief Converters from scikit-learn model. 

5""" 

6import numpy 

7from .g_sklearn_type_helpers import check_type 

8from .grammar.gactions import MLActionVar, MLActionCst, MLActionReturn 

9from .grammar.gactions_tensor import MLActionTensorDiv, MLActionTensorSub 

10from .grammar.gmlactions import MLModel 

11 

12 

13def sklearn_standard_scaler(model, input_names=None, output_names=None, **kwargs): 

14 """ 

15 Converts a `standard scaler <http://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.StandardScaler.html>`_ 

16 model into a *grammar* model (semantic graph representation). 

17 

18 @param model scikit-learn model 

19 @param input_names name of the input features 

20 @param output_names name of the output predictions 

21 @param kwargs additional parameters (none) 

22 @return graph model 

23 

24 If *input* is None or *output* is None, default values 

25 will be given to the outputs 

26 ``['Prediction', 'Score']`` for the outputs. 

27 If *input_names* is None, it wil be ``'Features'``. 

28 

29 No additional parameters is considered. 

30 """ 

31 if output_names is None: 

32 output_names = ['Prediction', 'Score'] 

33 if input_names is None: 

34 input_names = 'Features' 

35 

36 from sklearn.preprocessing import StandardScaler 

37 check_type(model, StandardScaler) 

38 

39 lmean = MLActionCst(model.mean_.ravel().astype(numpy.float32)) 

40 lscale = MLActionCst(model.scale_.ravel().astype(numpy.float32)) 

41 

42 lvar = MLActionVar(model.var_.astype(numpy.float32), input_names) 

43 lno = MLActionTensorSub(lvar, lmean) 

44 lno = MLActionTensorDiv(lno, lscale) 

45 ret = MLActionReturn(lno) 

46 return MLModel(ret, output_names, name=StandardScaler.__name__)