Add a function to load real-world datasets

[kunwuz]

Load real-world datasets. Processed datasets are from https://github.com/cmu-phil/example-causal-datasets/tree/main.

Description:

data, labels = load_dataset(dataset_name)

Parameters

dataset_name : str, ['sachs', 'boston_housing', 'airfoil']

Returns

data = np.array
labels = list

Usage example:

from causallearn.utils.Dataset import load_dataset

data, labels = load_dataset('sachs')

Output:

[[2.640e+01 1.320e+01 8.820e+00 ... 1.700e+01 4.490e+01 4.000e+01]
[3.590e+01 1.650e+01 1.230e+01 ... 3.370e+00 1.650e+01 6.150e+01]
[5.940e+01 4.410e+01 1.460e+01 ... 1.140e+01 3.190e+01 1.950e+01]
...
[2.810e+01 4.490e+00 1.880e+01 ... 1.000e+00 1.750e+00 2.000e+00]
[3.460e+01 7.100e+00 5.730e+00 ... 4.450e+01 1.382e+03 2.440e+00]
[3.050e+01 1.010e+00 7.300e+00 ... 1.000e+00 1.000e+00 1.650e+00]]
['raf', 'mek', 'plc', 'pip2', 'pip3', 'erk', 'akt', 'pka', 'pkc', 'p38', 'jnk']