Automatically parallelize your for comprehensions at compile time.
// build.sbt
libraryDependencies ++= Seq(
"io.github.kitlangton" %% "parallel-for" % "0.0.6",
"io.github.kitlangton" %% "parallel-for-zio" % "0.0.6", // for ZIO 2
"io.github.kitlangton" %% "parallel-for-zio1" % "0.0.6", // for ZIO 1
"io.github.kitlangton" %% "parallel-for-zio-query" % "0.0.6", // for ZIO QUERY (for ZIO 2)
"io.github.kitlangton" %% "parallel-for-cats-effect-3" % "0.0.6", // for Cats Effect 3
)
Wrapping a for-comprehension in par
will automatically parallelize it at compile time.
import parallelfor._
import parallelfor.interop.zio._ // Change this line depending on your effect system
import zio._
val program =
par {
for {
users <- loadUsers
files <- loadFiles
config <- loadConfig
result <- process(users, files, config)
_ <- fireTheMissiles
} yield result
}
The par
macro will rewrite the above program to—essentially—the following:
val program =
for {
(users, files, config, _) <- loadUsers zipPar loadFiles zipPar loadConfig zipPar fireTheMissiles
result <- process(users, files, config)
} yield result
Neato!
Similar project: VirtusLab/avocADO (only for Scala 3)
- Work with ZIO and ZManaged
- Maximize parallelization with topological sorting
- Get it to work with val assignments inside of for-comprehensions
- Set up publishing
- Generalize to work with any zippable structure (implement against
Parallelizable
type-class) - Fix withFilter interaction
- Cross-build for JS and Native
- Cross-build against 2.11/2.12
- Get it to work with Scala 3