Version 1.0.0: Submodels

With the imminent release of another update for Voyageur, Improv is mature enough that it seems like time to bump the version number up to 1.0. A major new feature has been added: Submodels.

Submodels are a way of creating complex models with nested underlying entities. You could represent individual characters in a procedurally-generated story, for example, and have Improv generate text for them consistently and independently across several characters:

Snippet	Result
"Let's go," [>protagonist:says_desc]	"Let's go," she said gruffly.
"I'm right behind you," [>loveInterest:says_desc]	"I'm right behind you," he said.
"Not so fast," [>antagonist:says_desc]	"Not so fast," they intoned.
"Aieeee!", [>loveInterest:shouted]	"Aieeee!", he screamed!

Check out the documentation for more info on submodels.

Work on this release was possible thanks to the support of Patreon supporters. Special thanks to Emily Short, Kevin Snow, Liza Daly, and Doug Orleans.

Version 0.6

This is the last feature release in a while, I promise.

New in this version:

• Binding; snippet objects can now have a property bind: true which causes that generator to have its output "frozen" for a given model.
• Chained directives: Templates can now call methods on model objects to process text.
• Built-in chains: Automatic capitalisation and a/an.

Version 0.5

This is the first quasi-stable release of Improv; I don't necessarily foresee drastic API changes in the near future any more, but I also don't guarantee none will happen until version 1.0. This version should be adequately documented and reasonably free of bugs.

New in this release:

• Proper documentation which can be viewed at Read the Docs.
• The new salienceFormula option can be used to decide how strict Improv will be about selecting phrases.

As usual you can get this from npm.