Error in documentation. Multidimensional grid_points for evaluationg FFTKDE needs to be numpy array.

[teakfi]

import seaborn as sns
from KDEpy import FFTKDE
import numpy as np

penguins = sns.load_dataset("penguins")
chinstrap = penguins[penguins["species"]=="Chinstrap"].iloc[:,2:6].dropna()

fit=FFTKDE().fit(chinstrap.to_numpy())

lengthspace = np.linspace(25,65,5)
depthspace = np.linspace(10,25,5)
flipperspace = np.linspace(150,250,5)
massspace = np.linspace(2000,8000,5)

def gridcreator1(space1,space2,space3,space4):
    grid = []
    for s1 in space1:
        for s2 in space2:
            for s3 in space3:
                for s4 in space4:
                    row = np.array([s1,s2,s3,s4])
                    grid.append(row)
    return grid

testgrid = gridcreator1(lengthspace,depthspace,flipperspace,massspace)

result = fit.evaluate(testgrid)

produces error:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[9], line 27
     23     return grid
     25 testgrid = gridcreator1(lengthspace,depthspace,flipperspace,massspace)
---> 27 result = fit.evaluate(testgrid)

File E:\anaconda3\envs\analysis\Lib\site-packages\KDEpy\FFTKDE.py:145, in FFTKDE.evaluate(self, grid_points)
    143 # Extra verification for FFTKDE (checking the sorting property)
    144 if not grid_is_sorted(self.grid_points):
--> 145     raise ValueError("The grid must be sorted.")
    147 if isinstance(self.bw, numbers.Number) and self.bw > 0:
    148     bw = self.bw

ValueError: The grid must be sorted.

casting testgrid as np array fixes this

result = fit.evaluate(np.asarray(testgrid))

works as intended (I hope).

[tommyod]

Thanks for reporting the issue. A PR on this is welcome :)