Quick guide to create a custom ChatterBot corpus

The ChatterBot python library is a great introduction to machine learning. It is very easy to create and train your own custom data by creating a YAML file. Luckily, ChatterBot has a custom corpus file and folder where you can add the content you want for the chat bot.

Naming a category

You want to start off naming the corpus category. It can be any topic you want.

categories:
- Hello

Writing dialog

Added dialog is important if you want the chat bot to reply with a specific answer. Your input is denoted by - - while the response for the bot is starts with - . After when you are done save the file in the custom folder. You can also add multiple YAML files to the custom folder.

conversations: 
# Your input
- - hello
# Bot response
-   world

The YAML file should look like this:

categories:
- Hello
conversations: 
- - hello
  - world

Training the custom file

You can now train the bot with the custom corpus that you created.

trainer = ChatterBotCorpusTrainer(bot)
trainer.train('chatterbot.corpus.custom')

The bot can be trained with multiple files within a folder. For example, we have a folder with various YAML files.

trainer.train('chatterbot.corpus.customFolder')

That's it for the quick introduction. For more information visit ChatterBot for documentation.