carefree0910/carefree-data

A data processing module implemented with numpy

carefree-data's Issues

carefree-data and carefree-learn appear to require numpy == 1.20.0

When importing either of:

from cfdata.tabular import TabularDataset
import cflearn

When the following lines in cython_wrappers.py is run with numpy==1.19.2 it will fail

try:
    from .cython_utils import *
except ImportError:
    raise

# with error
ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 88 from C header, got 80 from PyObject

numpy 1.20 is not yet available in anaconda, but if numpy is uninstalled & numpy==1.20.0 is specifically installed via pip, the error disappears.

Loading Titanic Dataset

I want to try use your program on Titanic but am unsure how to split the csv file into x and y. I tried the code below but it did not work:
train = 'titanic_train.csv'
label_column = 'Survived'
data = TabularData().read(train)
assert data.processed == data.transform(train)
x, y = data.processed.xy