Update app.py

app.py

@@ -123,7 +123,7 @@ if task == "Clustering":
         ('cluster', KMeans(n_clusters = n_clusters, n_init = n_init, max_iter = max_iter)),
     ])
 
-@st.cache
+@st.cache(allow_output_mutation=True)
 def fit_text_clf(X, y):
     text_clf.fit(X, y)
     return text_clf
@@ -154,10 +154,16 @@ form_explainer.form_submit_button("Submit")
 
 te = TextExplainer(random_state=42, char_based=char_based, n_samples = number_samples, position_dependent=position_dep)
 
+@st.cache(allow_output_mutation=True)
+def fit_text_explainer(X, predict_proba):
+    te.fit(X, predict_proba)
+    return te
+
 input_choice = st.checkbox("Check this if you want to enter your own example to explain", value = False)
 if input_choice == False:
     record_to_explain = st.number_input("Enter the index of the document from the original dataset to interpret", value = 30)
-    te.fit(df[column_name][record_to_explain], text_clf.predict_proba)
+    #te.fit(df[column_name][record_to_explain], text_clf.predict_proba)
+    fit_text_explainer(df[column_name][record_to_explain], text_clf.predict_proba)
     if task == "Classification":
         st.write("Ground truth label")
         st.write(df[labels_column_name][record_to_explain])
@@ -172,7 +178,8 @@ if input_choice == False:
         st.write(model_prediction)
 else:
     record_to_explain = st.text_area("Enter the example document to explain", value = text_example)
-    te.fit(record_to_explain, text_clf.predict_proba)
+    #te.fit(record_to_explain, text_clf.predict_proba)
+    fit_text_explainer(record_to_explain, text_clf.predict_proba)
     if task == "Classification":
         st.write("Model prediction")
         model_prediction = text_clf.predict([record_to_explain])