chatbot.py

# The Shani Sir chatbot

from chatterbot import ChatBot
from chatterbot.comparisons import JaccardSimilarity

shanisirbot = ChatBot('The Shani Sir Bot',
                      storage_adapter='chatterbot.storage.SQLStorageAdapter',
                      logic_adapters=[{"import_path": "chatterbot.logic.BestMatch",
                                       "statement_comparison_function": JaccardSimilarity,
                                       'maximum_similarity_threshold': 0.99}],
                      preprocessors=['chatterbot.preprocessors.clean_whitespace'],
                      database_uri='sqlite:///db.sqlite3',
                      read_only=False)  # Set to True to disable further learning from conversations

# NOTE: response_selection_method can also be get_random_response or get_most_frequent_response

shanisirbot.initialize()  # Does any work that needs to be done before the chatbot can process responses.
get_tags = shanisirbot.storage.tagger.get_bigram_pair_string

# def train_with(corpus: str) -> None:
#     """
#     Trains the bot using the specified corpus
#     eng ---> chatterbot.corpus.english (standard English corpus from chatterbot_corpora)
#     woz ---> ./MULTIWOZ2.1 (Multi-Domain Wizard-of-Oz dataset from http://dialogue.mi.eng.cam.ac.uk/index.php/corpus/)
#     ubu ---> Will download and extract the Ubuntu dialog corpus if that has not already been done.
#     """
#
#     from chatterbot.trainers import ChatterBotCorpusTrainer, UbuntuCorpusTrainer
#     import time
#
#     if corpus == 'ubu':  # WARNING: TAKES A REALLY LONG TIME
#         start = time.time()  # (TOOK 114000 SECONDS = 31 HRS TO EXTRACT & TRAIN FOR UNCLE SAM, NOT INCLUDING DL TIME)
#         corpus_trainer = UbuntuCorpusTrainer(shanisirbot)
#         corpus_trainer.train()
#     else:
#         start = time.time()
#         corpus_trainer = ChatterBotCorpusTrainer(shanisirbot)
#         if corpus == 'eng':
#             corpus_trainer.train("chatterbot.corpus.english")
#         elif corpus == 'woz':
#             corpus_trainer.train('./data/MULTIWOZ2.1/attraction_db.json',
#                                  './data/MULTIWOZ2.1/data.json',
#                                  './data/MULTIWOZ2.1/dialogue_acts.json',
#                                  './data/MULTIWOZ2.1/hospital_db.json',
#                                  './data/MULTIWOZ2.1/hotel_db.json',
#                                  './data/MULTIWOZ2.1/ontology.json',
#                                  './data/MULTIWOZ2.1/police_db.json'
#                                  './data/MULTIWOZ2.1/restaurant_db.json',
#                                  './data/MULTIWOZ2.1/taxi_db.json',
#                                  './data/MULTIWOZ2.1/testListFile.json',
#                                  './data/MULTIWOZ2.1/train_db.json',
#                                  './data/MULTIWOZ2.1/valListFile.json')
#         else:
#             print("Invalid corpus.")
#             return
#     end = time.time()
#     time_taken = end - start
#     print(f"\n\nThe Shani Sir chatbot has been trained using the corpus {corpus}. Time taken: {time_taken}s")