Чат-бот с NLTK с типом запроса и параметрами

Question

Я новичок в NLTK и пытаюсь создать чат-бот с типом запроса и параметром.

Например,

      corpus = [
        {
            name: "appt-count",
            text: "How many appointments I have for today?"
        },
        {
            name: "appt-count",
            text: "What is my total appointments for today?"
        },
        {
            name: "appt-list",
            text: "What are all my appointments today?"
        },
        {
            name: "appt-list",
            text: "Call you tell me my appointments today?"
        },
        {
            name: "appt-view",
            text: "What is my next appointment?"
        },
        {
            name: "appt-view",
            text: "What is my appointment after lunch?"
        },
        {
            name: "appt-view",
            text: "What is my first appointment today?"
        },
        {
            name: "appt-view",
            text: "What is my last appointment today?"
        }
    ]

Here, When user input text, it is supposed to return the corresponding name, so that the system will invoke the corresponding API and return the result to the user.

Когда пользователь говорит: «Сколько у меня встреч на сегодня?» предполагается, что он возвращает «количество приложений», а также параметр «сегодня», чтобы система проверяла сегодняшний день, и пользователь мог также ввести здесь некоторую дату.

Я пытаюсь использовать LogisticRegression из sklearn.linear_model. Получение ошибки в коде np.array. Я на правильном пути?

      import nltk
import pickle
import pandas as pd
import numpy as np
from nltk.stem import PorterStemmer, WordNetLemmatizer
from nltk.tokenize import sent_tokenize, word_tokenize
from nltk.classify import ClassifierI

from sklearn.linear_model import LogisticRegression, SGDClassifier
from sklearn.feature_extraction.text import CountVectorizer,TfidfVectorizer

training = []
ps = PorterStemmer()

df=pd.read_json("./train_data.json", orient="records")

for name, text in zip(df['name'], df['text']):
  words = word_tokenize(text)
  stemWord = [ps.stem(w.lower()) for w in words]
  training.append([stemWord , name])

// Error:- Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.
training = np.array(training) 
train_x = list(training[:,0])
train_y = list(training[:,1])

print("Training data created")

model = LogisticRegression()

model=model.fit(train_x,train_y)