[BUG][Experimental] AttributeError: 'numpy.ndarray' object has no attribute 'to_numpy'

[sbobek]

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In[5], line 6
      3 from lux.pyuid3.data import Data
      4 data = Data.parse_dataframe(test[features+[target]])
----> 6 predictions = [int(x.get_name()) for x in lux.uid3.predict(data.instances)]
      7 print(f'Accuracy : {accuracy_score(predictions, test[target])}')

File ~/Software/lux/src/lux/pyuid3/uid3.py:554, in UId3.predict(self, X)
    552 def predict(self, X: pd.DataFrame):
    553     predictions = []
--> 554     for instance in X.to_numpy():
    555         att_stats = self.tree.predict(instance, list(X.columns))
    556         predictions.append(att_stats.get_most_probable())

AttributeError: 'numpy.ndarray' object has no attribute 'to_numpy'

Code to reproduce:

from lux.lux import LUX
from sklearn import datasets
from sklearn.model_selection import train_test_split
import numpy as np
import pandas as pd
from matplotlib import pyplot as plt
import seaborn as sns
from sklearn.ensemble import RandomForestClassifier
wine = datasets.load_wine()


features = [f.replace('/','_') for  f in wine['feature_names']]
target = 'class'
rs=42
fraction=0.1

#create daatframe with columns names as strings (LUX accepts only DataFrames with string columns names)
df_wine = pd.DataFrame(wine.data,columns=features)
df_wine[target] = wine.target

#train classifier
train, test = train_test_split(df_wine, random_state=rs)
clf = RandomForestClassifier(random_state=42)#svm.SVC(probability=True, random_state=rs)
clf.fit(train[features],train[target])
clf.score(test[features],test[target])
## Without predictor uncertainty
#pick some instance from datasetr
i2e = train[features].sample(1, random_state=42).values
i2e

#train lux on neighbourhood equal 30% instances
lux = LUX(predict_proba = lambda x: np.round(clf.predict_proba(x)), neighborhood_size=int(len(train)*fraction),max_depth=2,  node_size_limit = 1, grow_confidence_threshold = 0 )
lux.fit(train[features], train[target], instance_to_explain=i2e,class_names=[0,1,2])

#see the justification of the instance being classified for a given class
lux.justify(np.array(i2e))
#calculate decision tree score over this dataset
from sklearn.metrics import accuracy_score, f1_score, precision_score, recall_score
from lux.pyuid3.data import Data
data = Data.parse_dataframe(test[features+[target]])
    
predictions = [int(x.get_name()) for x in lux.uid3.predict(data.instances)]
print(f'Accuracy : {accuracy_score(predictions, test[target])}')