0.3.1

Adds official support for Django 5.1.

0.3.0

What's Changed

• Feature: add new corpus entries to an existing model by @andrlik in #14

Full Changelog: 0.2.4...0.3.0

0.2.4

Adds the ability to optionally change the state size of text models using MARKOV_STATE_SIZE in your settings.py.

0.2.3

Full Changelog: 0.2.2...0.2.3

• Adds a cached property, is_compiled_model to MarkovTextModel so a user doesn't need to know markovify internals to check.
• Adds a possible exception MarkovEmptyError raised by is_compiled_model if the model does not yet have any data.

0.2.2

Full Changelog: 0.2.1...0.2.2

• Fixes issue where CORPUS_TEXT_MAX_CHARS_LIMIT is not set in a more graceful manner, falling back to a default value of 0 (no limit).

0.2.1

What's Changed

• Fix license detection by @andrlik in #5
• Add testing runs for django4.2 by @andrlik in #6

New Contributors

• @andrlik made their first contribution in #5

Full Changelog: 0.2.0...0.2.1

0.2.0

Full Changelog: 0.1.1...0.2.0

• Improves sentence generation performance by loading the text model into a cached property.
  • Cached properties are cleared when calling refresh_from_db.
• Enables storing non-compiled models to the database for use in combining multiple models into a single chain.
  • Adds boolean settings parameter MARKOV_STORE_COMPILED_MODELS to set default behavior.
  • This can be overridden when calling update_model_from_corpus using the store_compiled kwarg.
  • Adds ability to auto-detect if a loaded model is compiled or not, and will automatically compile for sentence generation if needed.
• Adds support for the tries directive when calling generate_sentence.
• Adds support for combining multiple models into one via MarkovTextModel.combine_models

0.1.1

• Fix project metadata
• Move extraneous dependencies to dev-dependencies.

v0.1.0

Initial Release 🎉