detailed summary type does not work: Callback error updating interaction-summary-graph

in explainerdashboard version 0.34

when using the detailed summary type on the XGBRegressionExplainer i get the following error (in dash):

Callback error updating interaction-summary-graph-60.figure, interaction-summary-index-col-60.style 17:56:37 Callback error updating interaction-summary-graph-60.figure, interaction-summary-index-col-60.style

code to replicate:

from sklearn.datasets import load_diabetes
import pandas as pd

data = load_diabetes()
X = pd.DataFrame(data = data['data'],columns = data['feature_names'])
y = data['target']

from xgboost import XGBRegressor

sklearnxgbr = XGBRegressor()
sklearnxgbr.fit(X,y)

from explainerdashboard.explainers import XGBRegressionExplainer
from explainerdashboard import ExplainerDashboard

explainer = XGBRegressionExplainer(sklearnxgbr, X,y)

ExplainerDashboard(explainer, 
                  summary_type='detailed', 
                  ).run()

All the other plots do work, including the aggregate summary plot but not the detailed (compact_dot) feature dependency and interaction summary plots

[oegedijk]

Ah, found it! There was a bug when you don’t define any onehot encoded columns. Fixed it, will try to get a new release out tomorrow.

So this fails:

from sklearn.ensemble import RandomForestClassifier
from explainerdashboard import ClassifierExplainer, ExplainerDashboard
from explainerdashboard.datasets import titanic_survive, titanic_names

X_train, y_train, X_test, y_test = titanic_survive()
explainer = ClassifierExplainer(RandomForestClassifier().fit(X_train, y_train), X_test, y_test)
ExplainerDashboard(explainer).run()

but this should work:

from sklearn.ensemble import RandomForestClassifier
from explainerdashboard import ClassifierExplainer, ExplainerDashboard
from explainerdashboard.datasets import titanic_survive, titanic_names

X_train, y_train, X_test, y_test = titanic_survive()
explainer = ClassifierExplainer(RandomForestClassifier().fit(X_train, y_train), X_test, y_test, 
                                cats=['Sex', 'Deck', 'Embarked'])
ExplainerDashboard(explainer).run()

[hugocool]

Hi @oegedijk , it definitely works!

many thanks!