TwitchMarkovChain Versions Save

Full Changelog:

• Improve detokenization performance with more recent NLTK versions (f994465f0f5c9304ebb9830926f3df130cf9643a, #31)

• Fix bug with contractions not learning correctly (#29)

• Resolved No such table error issue on first execution of the Bot. (#23)

• Tom Aarsen

v2.1 update log:

• Added GenerateCommands field - Specifies the generation commands that the bot will listen for. Defaults to ["!generate", "!g"]. Useful if your chat is used to commands with ~, -, /, etc.
• Added AllowGenerateParams field - Specifies whether the chat is allowed to supply a partial sentence which the bot finishes, e.g. !generate hello, I am. If false, all values after the generation command will be ignored.
• Place blacklist.txt on .gitignore.

See #18 and #19.

Thanks to @justinrusso for the help.

• Tom Aarsen

v2.0 update log:

• Punctuation (commas, dots, apostrophes, etc.) are now split as a separate word in the database.
• Added AllowedUsers field - gives developer (or whichever users are specified here) the permission to bypass the cooldown, or use the generate command even when the command is disabled.
• Added WhisperCooldown field - whether the Bot whispers the cooldown message. Otherwise, do not send the cooldown.
  • Still always whisper in certain situations, such as when the broadcaster updates the cooldown by whispering the Bot.
• Added EnableGenerateCommand field - enables or disables !generate. Broadcaster and AllowerUsers can override this setting, and still generate.
• Start a second sentence if the generated sentence is too short, uses a new MinSentenceWordAmount value in settings.json.
• Added a SentenceSeparator value, which is placed inbetween sentences, when multiple sentences are generated (only when the first sentence was too short according to MinSentenceWordAmount)
• Renamed settings.txt with settings.json, which is automatically updated from v1.0
• Update the database according to the punctuation changes. This creates a ..._backup.db file with the old database.
• Add a Tokenizer file that handles splitting up a sentence into tokens, and merging tokens back into sentences.
• Improved README - removed troublesome HTML tags, improved readability, and added new fields in settings.
• Modified default outputs slightly (e.g. after using !enable).
• The default help Timer is now once every 5 hours, instead of off.
• Added .gitignore.
• Added typing to parameters in methods.
• Added better docstrings to methods.
• Improved code quality.

See #17, #14 and #16.

Thanks to @justinrusso for the help.

• Tom Aarsen

Full stable release.