Calyx

Generative text processing for C# and Unity applications.

Build

dotnet build

Tests

dotnet test

Install

Unity

Create a new Unity project or open an existing project.

In the Assets directory of your project, create a new directory called Libraries (or adapt to suit your preferred style of organising).

Download Calyx.dll from the compiled resources under Build/Release on eng-git.

Move Calyx.dll into the Assets/Libraries path of your Unity project.

Create a new C# Script in your Unity project and assign it to a new empty game object (or add it to an existing game object in your scene).

Add the following grammar script to the Awake() or Start() hook of your script.

using UnityEngine;
using Calyx;

public class TextSample : MonoBehaviour
{
  void Start()
  {
    Grammar hello = new Grammar(P => {
      P.Start("Hello from Unity");
    });

    Debug.Log(hello.Generate().text);
  }
}

Run the scene in the Unity editor and you should see the phrase Hello from Unity appear in the console.

Usage

Configuring the PRNG

Requires an instance of System.Random not the static class UnityEngine.Random.

If you have a predefined integer to use as a seed value, pass it in using the Seed option:

int aSeed = 72365472659210;

Grammar grammar = new Grammar(Seed: aSeed);

You can also pass in an instance of Random with the Rng option.

int bSeed = 82673452385460;

Grammar grammar = new Grammar(
    Rng: new System.Random(bSeed)
);

Defining grammar rules

Use the object delegate constructor to return a preconfigured instance with all the rules defined:

Grammar plums = new Grammar(P => {
    P.Start("I have {eaten} the {plums}");
    P.Rule("eaten", new[] { "eaten", "taken", "devoured" });
    P.Rule("plums", new[] { "plums", "apricots", "pears" });
});

Rule definitions support fluent interface conventions, so can also be constructed as a single statement:

Grammar plums = new Grammar(P => {
    P.Start("I have {eaten} the {plums}")
        .Rule("eaten", new[] { "eaten", "taken", "devoured" })
        .Rule("plums", new[] { "plums", "apricots", "pears" });
});

The Start and Rule methods are overloaded, supporting both arrays of strings, and single strings:

Grammar plums = new Grammar(P => {
    P.Rule("single", "I have eaten the plums")
     .Rule("alternate", new[] {
        "I have eaten the plums",
        "That were in the icebox"
    });
});

Generating Results

Call Generate on the grammar instance to get a result object from a unique run of the grammar and use the Text property of the result to access the generated text.

Result result = plums.Generate();

Console.WriteLine(result.Text);

By default, grammars will generate top-down from the start rule. To start from an alternative rule in the grammar, pass the start symbol as a string to Generate:

Result result = plums.Generate("alternate");

Console.WriteLine(result.Text);

To append template context rules at runtime, you can pass a transient block of rules to the grammar as part of the call to Generate. This can be useful if you want to mix data from different life cycles of your program state but (presumably—we haven’t performance tested this yet!) will make grammar generation slower though in most cases shouldn’t be noticeable.

Result result = plums.Generate(new Dictionary<string, string[]>() {
  ["single"] = new[] { "I have eaten the plums" }
  ["alternate"] = {
    "I have eaten the plums",
    "That were in the icebox"
  }
});

Console.WriteLine(result.Text);

Other Features

Not all the features are ported over yet and the template syntax is not yet documented but you can read about the general format and structure of the language design in the original SYNTAX document.

Roadmap

Rough plan for stabilising the API and features for a 1.0 release.

Version	Features planned
0.1	block constructors and basic grammar features
0.2	runtime generator context in template substitution
0.3	memo nodes and expression syntax
0.4	uniq nodes and expression syntax
0.5	weighted branch productions and definition parser
0.6	modifying filter chains and expression syntax
0.7	affix tree productions, definition parser and expression syntax
0.8	error handling and stack trace diagnostics
0.9	Unity JSON affordances and file format integration
0.10	Parity check with JavaScript and Ruby engines
0.11	API documentation and code comments
0.12	NuGet and Unity package manager nonsense