Value Error: y contains previously unseen labels:

I've used Decision Tree Classifier and I want to enter my input as a string rather than giving an integer value, but it gives me error like:

Traceback (most recent call last):
  File "D:/backup code for odoo project/New folder/New folder/main.py", line 38, in <module>
    theme_res = lebel_encoder.transform(theme_input)
  File "C:\Users\Dell\AppData\Local\Programs\Python\Python38\lib\site-packages\sklearn\preprocessing\_label.py", line 277, in transform
    _, y = _encode(y, uniques=self.classes_, encode=True)
  File "C:\Users\Dell\AppData\Local\Programs\Python\Python38\lib\site-packages\sklearn\preprocessing\_label.py", line 121, in _encode
    return _encode_numpy(values, uniques, encode,
  File "C:\Users\Dell\AppData\Local\Programs\Python\Python38\lib\site-packages\sklearn\preprocessing\_label.py", line 50, in _encode_numpy
    raise ValueError("y contains previously unseen labels: %s"
ValueError: y contains previously unseen labels: ['Food', 'cafe', 'sticky'] 

Code:

import pandas as pd
from sklearn.preprocessing import LabelEncoder
from sklearn import tree
df = pd.read_csv("new_data.csv", encoding='latin1')

inputs = df.drop('selected_theme', axis='columns')
target = df['selected_theme']

lebel_encoder = LabelEncoder()

inputs['main_cat_n'] = lebel_encoder.fit_transform(inputs['main_cat'])
inputs['sub_cat_n'] = lebel_encoder.fit_transform(inputs['sub_cat'])
inputs['nav_bar_n'] = lebel_encoder.fit_transform(inputs['nav_bar'])

inputs_n = inputs.drop(['main_cat', 'sub_cat', 'nav_bar'], axis='columns')

 model = tree.DecisionTreeClassifier()
 model.fit(inputs_n, target)
 print(model.score(inputs_n, target))

 theme_input = ['Food', 'cafe', 'sticky']
 theme_res = lebel_encoder.transform(theme_input)
 result_theme = model.predict(theme_res)
 print(result_theme)

The error occurs before classifier, it happens on this line

 theme_res = lebel_encoder.transform(theme_input)

error message says to you that your label_encoder never saw such categories as "Food", "cafe", "sticky". It happens because you rewrite your LabelEncoders. You should use separate LabelEncoders for different features, e.g.:

categorical_features = ['main_cat', 'sub_cat', 'nav_bar']
encoders = dict()
for cat in categorical_features:
    encoders[cat] = LabelEncoder()
    inputs[f'{cat}_n'] = encoders[cat].fit_transform(inputs[cat])

inputs_n = inputs.drop(['main_cat', 'sub_cat', 'nav_bar'], axis='columns')
...