Classes/structs, pattern matching, safe counters... and more!
Your one-stop library for programming tools not already in core Typst.

Pronounced ‘tipsy’, because I think that’s funny and it’s a nice pun on ‘Typst’. 😄

Provides tools for programming geeks:

• classes (i.e. structs, custom types)
• pattern matching
• enums
• safe counters (no need to choose a unique string)
• trees-of-counters (i.e. subcounters)
• string formatting
• namespaces of objects that can be mutually referential
• runtime type checking

Installation

Typst will autodownload packages on import:

#import "@preview/typsy:0.1.0"

What’s in the box?

Classes

Classes with fields and methods:

#import "@preview/typsy:0.1.0": class

#let Adder = class(
    fields: (x: int),
    methods: (
        add: (self, y) => {self.x + y}
    )
)
#let add_three = (Adder.new)(x: 3)
#let five = (add_three.add)(2)

Pattern-matching

Simple type checking:

#import "@preview/typsy:0.1.0": Array, Int, matches

// Fixed-length case.
#matches(Array(Int, Int), (3, 4)) // true
// Variable-length case.
#matches(Array(..Int), (3, 4, 5, "not an int")) // false

More complicated match-case statements:

#import "@preview/typsy:0.1.0": Arguments, Int, Str, case, match, matches

// Option 1: if/else
#let fn-with-multiple-signatures(..args) = {
    if matches(Arguments(Int), args) {
        // ...
    } else if matches(Arguments(Str), args) {
        // ...
    } else if matches(Arguments(Str, level: Int), args) {
        // ...
    } else {
        panic
    }
}

// Option 2: match/case
#let fn-with-multiple-signatures(..args) = {
    match(args,
        case(Arguments(Int), ()=>{
            // ...
        }),
        case(Arguments(Str), ()=>{
            // ...
        }),
        case(Arguments(Str, level: Int), ()=>{
            // ...
        }),
    )
}

Observe the capitalisation. All patterns are capitalised to distinguish them from their corresponding type.

Enums

Also using the same pattern-matching capabilities as above:

#import "@preview/typsy:0.1.0": case, class, enumeration, match

#let Shape = enumeration(
    Rectangle: class(fields: (height: int, width: int)),
    Circle: class(fields: (radius: int)),
)
#let area(x) = {
    match(x,
        case(Shape.Rectangle, ()=>{
            x.height * x.width
        }),
        case(Shape.Circle, ()=>{
            calc.pi * calc.pow(x.radius, 2)
        }),
    )
}

Safe counters

Counters without needing to cross your fingers and hope that you’re using a unique string each time:

#import "@preview/typsy:0.1.0": safe-counter

#let my-counter1 = safe-counter(()=>{})
#let my-counter2 = safe-counter(()=>{})
// ...these are different counters!
// (All anonymous functions have different identities to the compiler.)

Tree counters / subcounters

Create trees of counters, including using existing counters as starting points. This is particularly useful for creating theorem counters that increment with the heading.

#import "@preview/typsy:0.1.0": tree-counter

// Set up counters
#let heading-counter = tree-counter(heading, level: 1)
#let theorem-counter = (heading-counter.subcounter)(()=>{})  // uses safe-counter internally!
#let corollary-counter = (theorem-counter.subcounter)(()=>{})

// Usage
#set heading(numbering: "1")
#let theorem(doc) = [Theorem #(theorem-counter.take)(): #doc]
#let corollary(doc) = [Corollary #(corollary-counter.take)(): #doc]
= First Section
#theorem[Let ...]  // Theorem 1.1: Let ...
#theorem[Let ...]  // Theorem 1.2: Let ...
#corollary[Let ...] // Corollary 1.2.1: Let ...
= Second Section
#theorem[Let ...] // Theorem 2.1: Let ...

String formatting

Rust-like string formatting:

#import "@preview/typsy:0.1.0": fmt, panic-fmt

#let msg = fmt("Invalid input `{}`, expected `{}`.", foo, bar)

// shorthand for `panic(fmt(...))`
#panic-fmt("Invalid input `{}`, expected `{}`.", foo, bar)

Runtime type checking

Wrap functions to check their inputs and outputs. This builds on top of the pattern-matching capablities above.

#import "@preview/typsy:0.1.0": Arguments, typecheck

#let add_integers = typecheck(Arguments(Int, Int), Int, (x, y) => x + y)
#let five = add_integers(2, 3)  // ok
#add_integers("hello ", "world")  // panic!

Namespaces of mutually-referential objects

Build a namespace by providing lambda functions which return their object. Access any object in a namespace via ns(object-name):

#import "@preview/typsy:0.1.0": namespace

#let ns = namespace(
    foo: ns => {
        let foo(x) = if x == 0 {"FOO"} else {ns("bar")(x - 1)}
        foo
    },
    bar: ns => {
        let bar(x) = if x == 0 {"BAR"} else {ns("foo")(x - 1)}
        bar
    },
)
#let foo = ns("foo")
#assert.eq(foo(3), "BAR")
#assert.eq(foo(4), "FOO")

For example, this can be used to implement mutually-recursive functions.

Documentation

All objects have detailed docstrings indicating their usage; see those for details.

The examples above demonstrate nearly every object in the public API. In addition to those above, the list of patterns that can be used for pattern-matching are:

Any, Arguments, Array, Bool, Bytes, Class, Content, Counter, Datetime, Decimal,
Dictionary, Duration, Float, Function, Int, Label, Literal, Location, Module,
Named, Never, None, Pos, Ratio, Refine, Regex, Selector, State, Str, Symbol,
Type, Union, Version

(They are capitalised to distinguish them from the underlying type.)

FAQ

Similar libraries:

• elembic offers a very different way to create custom classes.
• valkyrie offers object parsing that is somewhat similar to our type matching.
• headcount and rich-counters also offer tree counters. (Though I find our approach a bit simpler, and safer due to our ()=>{}-using safe counters.)
• oxifmt offers Rust-like string formatting. Theirs is far more developed and better than what we have; I just like avoiding dependencies.