Coverage for NeuralTSNE/NeuralTSNE/Utils/Preprocessing/preprocessing.py: 100%

1import numpy as np 

2import torch 

3 

4from NeuralTSNE.Utils.Preprocessing.Normalizers import normalize_columns 

5from NeuralTSNE.Utils.Preprocessing.Filters import ( 

6 filter_data_by_variance, 

7) 

8from NeuralTSNE.Utils.Writers.StatWriters import save_means_and_vars 

9 

10 

11def prepare_data(variance_threshold: float, data: np.ndarray) -> torch.Tensor: 

12 """ 

13 Prepare data for further analysis by filtering based on variance, 

14 saving means and variances, and normalizing columns. 

15 

16 Parameters 

17 ---------- 

18 `variance_threshold` : `float` 

19 Threshold for variance-based feature selection. 

20 `data` : `np.ndarray` 

21 Input data array. 

22 

23 Returns 

24 ------- 

25 `torch.Tensor` 

26 Processed and normalized data tensor. 

27 

28 Note 

29 ---- 

30 The function filters the input `data` based on the provided `variance threshold`, 

31 saves means and variances, and then normalizes the columns of the `data` before 

32 converting it into a `torch.Tensor`. 

33 """ 

34 filtered = filter_data_by_variance(data, variance_threshold) 

35 save_means_and_vars(data, filtered) 

36 if filtered is not None: 

37 data = filtered 

38 

39 data = torch.from_numpy(data).float() 

40 data = normalize_columns(data) 

41 return data