Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/polyvariant/named-functions

Micro-library for working with functions with named parameters.
https://github.com/polyvariant/named-functions

macros scala scala3

Last synced: about 1 month ago
JSON representation

Micro-library for working with functions with named parameters.

Awesome Lists containing this project

README 

          
# named-functions



Scala 3 macros for converting multi-parameter functions into functions with named parameters or named-tuple arguments, preserving the original parameter names.



## Quick overview



```scala

import namedfunctions.syntax.*



def greet(name: String, age: Int): String = s"$name is $age"



// namedTuple(...) — build a named tuple from local names

val age = 30

val name = "Alice"

val nt = namedTuple(name, age)



// .named — named parameters at call site

val f = greet.named

f(name = "Alice", age = 30) // "Alice is 30"



// .namedTupled — function from named tuple

val g = greet.namedTupled

g((name = "Alice", age = 30)) // "Alice is 30"



// .namedUntupled — reverse of tupling

val h: ((name: String, age: Int)) => String = t => s"${t.name} is ${t.age}"

h.namedUntupled(name = "Alice", age = 30) // "Alice is 30"



// .nameChecked — compile-time name validation, order-independent

val age = 30

val name = "Alice"

greet.nameChecked(age, name) // "Alice is 30" — reordered automatically

// greet.nameChecked(age, x) // compile error: unexpected: x; missing: name



// .applyProduct — apply from case class fields by name

case class Person(age: Int, name: String, email: String)

greet.applyProduct(Person(30, "Alice", "a@b.com")) // "Alice is 30" — extra fields ignored

```



## Installation



Available on Maven Central as `org.polyvariant::named-functions`.



In scala-cli:



```scala

//> using dep org.polyvariant::named-functions::

```



In sbt:



```scala

libraryDependencies += "org.polyvariant" %% "named-functions" % ""

```



In Mill:



```scala

ivy"org.polyvariant::named-functions:"

```



## Usage



```scala

import namedfunctions.syntax.*



def foo(entityId: Int, userId: String): Boolean = ???



val entityId = 1

val userId = "hello"

val params = namedTuple(entityId, userId)



// Wrap a function so that its parameters are named at the call site

val f = foo.named

f(entityId = 1, userId = "hello")



// Convert a function into a Function1 from a named tuple

val g = foo.namedTupled

g((entityId = 1, userId = "hello"))



// Convert a Function1 from a named tuple back into a named-parameter function

val h: ((entityId: Int, userId: String)) => Boolean = ???

val f2 = h.namedUntupled

f2(entityId = 1, userId = "hello")

```



### Multiple parameter lists



Methods with multiple parameter lists are supported — the result is a curried function with named parameters at each level:



```scala

def bar(entityId: Int)(userId: String): Boolean = ???



val f = bar.named

f(entityId = 1)(userId = "hello")



val g = bar.namedTupled

g((entityId = 1))((userId = "hello"))

```



### `NamedFunctions.of` / `NamedFunctions.apply` / `.named`



Converts a multi-parameter function `(a: A, b: B, ...) => R` into a function with named parameters, preserving the original parameter names.



### `NamedFunctions.tupled` / `.namedTupled`



Like `.tupled` but the resulting tuple type carries the parameter names from the original method. Converts a multi-parameter function into a `Function1` from a named tuple: `((a: A, b: B, ...)) => R`.



### `syntax.namedTuple`



Builds a named tuple directly from variable references or field accesses by using the final name segment as the label:



```scala

import namedfunctions.syntax.*



val a = 42

val b = "hello"



namedTuple(a, b) // (a = 42, b = "hello")

```



### `NamedFunctions.untupled` / `.namedUntupled`



The reverse of `tupled`. Converts a `Function1` from a named tuple into a multi-parameter function with named parameters.



### `.nameChecked`



Compile-time check that argument variable names exactly match the function's parameter names. Arguments are matched by name and automatically reordered, so order doesn't matter — only that the names are correct:



```scala

def foo(a: Int, b: String): String = s"$a-$b"



val a = 42

val b = "hello"



foo.nameChecked(a, b)    // compiles — "42-hello"

foo.nameChecked(b, a)    // also compiles — reordered to "42-hello"

foo.nameChecked(b, x)    // compile error: unexpected: x; missing: a

foo.nameChecked("hello") // compile error: requires variable references or field accesses

```



Arguments can be plain variable references or field accesses (e.g. `obj.field`). The last segment of the access is used as the name. Multiple parameter lists are supported — all arguments are passed flat:



```scala

def bar(entityId: Int)(userId: String): Boolean = ???



val entityId = 1

val userId = "hello"

bar.nameChecked(entityId, userId)

```



Field accesses work too — the field name is what matters:



```scala

case class Source(a: Int, b: String)

val src = Source(42, "hello")

foo.nameChecked(src.a, src.b) // compiles — "42-hello"

```



### `.applyProduct`



Applies a function using fields from a case class, matched by name (not position). The case class may have extra fields, but all function parameters must be present:



```scala

def foo(a: Int, b: String): String = s"$a-$b"



case class Params(b: String, a: Int)

foo.applyProduct(Params("hello", 42)) // "42-hello" — fields matched by name



case class Extended(a: Int, b: String, extra: Boolean)

foo.applyProduct(Extended(1, "hi", true)) // works — extra fields ignored

```



Multiple parameter lists are supported:



```scala

def bar(entityId: Int)(userId: String): Boolean = ???



case class Params(entityId: Int, userId: String)

bar.applyProduct(Params(1, "hello"))

```



## Chaining



Features can be composed — the output of one transformation is a valid input for another:



```scala

def foo(a: Int, b: String): String = s"$a-$b"



case class Params(b: String, a: Int)

foo.named.applyProduct(Params("hello", 42)) // "42-hello"



val a = 42

val b = "hello"

foo.named.nameChecked(a, b) // "42-hello"



// Round-trip through tupling and back

foo.namedTupled.namedUntupled.applyProduct(Params("hello", 42)) // "42-hello"

```



## Limitations



All features require the macro to extract parameter names from the call site's AST. This works with method references (`obj.method`), eta-expanded methods, and case class constructors (`Foo.apply`), but **not with function values stored in a `val`**:



```scala

val f = (a: Int, b: String) => s"$a-$b"

f.named          // compile error: Could not extract parameter names

f.nameChecked(a, b) // same

f.applyProduct(p)   // same

```



The parameter names exist in the lambda at the definition site, but are erased by the time the `val` reference reaches the macro.