Coroutines are a general control structure allowing flow control to pass cooperatively between two different routines. Coroutines in this library are computations which can suspend their execution and return control to their invoker, which can resume the computation. Coroutines can be used to implement pipelines as described in Coroutine Pipelines by Mario Blažević.
If you are creating asynchronous or concurrent pipelines you may be interested in:
If you need a featureful streaming library, you may also be interested in:
If you are looking for a FRP library, then you may be interested in:
coroutines with Spago:
spago install coroutines
This quick start briefly introduces the basics of
coroutines. For a more thorough, beginner-friendly introduction to the library, please see the library documentation.
The basic building block of
coroutines is the coroutine type
Co, which exhibits different behavior when it suspensd based on a functor
Emit ofunctor, the coroutine emits an output of type
oand is a
Await ifunctor, the coroutine waits for an input of type
iand is a
Transform i ofunctor, the coroutine waits for an input of type
i, and emits an output of type
o, and is a
A coroutine which emits can be thought of as a generator, where each yield produces a value. A coroutine which awaits can be thought of as an iteratee, where each yield demands a value. And a coroutine which transforms does both: it demands an input, and produces an output.
Using these three building blocks we can create some standard coroutines and form them into pipelines.
Here is a coroutine which generates natural numbers:
nats :: forall m. (Monad m) => Producer Int m Unit nats = go 0 where go i = do emit i go (i + 1)
The computation runs, emits a number, and suspends; when resumed it will emit the next number and then suspend again. It uses the
emit :: forall m o. Monad m => o -> Producer o m Unit
Here is a coroutine which accepts and prints strings:
printer :: forall a. Consumer String (Aff _) Unit printer = forever do s <- await lift (log s)
The computation suspends, awaiting a number. When it receives a number it logs the number and then suspends, awaiting a new number. It uses the
await :: forall m i. Monad m => Consumer i m i
Here is a coroutine which transforms inputs by showing them as strings:
showing :: forall a m. (Show a, Monad m) => Transformer a String m Unit showing = forever (transform show)
The computation suspends, awaiting an
a. When it receives the awaited value, it transforms it into a string, emits the new value, and suspends again, awaiting another
a. It uses the
transform :: forall m i o. Monad m => (i -> o) -> Transformer i o m Unit
These coroutines can be combined together using a handful of operators, the most common of which include:
connect :: forall o f m a. MonadRec m => Parallel f m => Producer o m a -> Consumer o m a -> Process m a infixr 2 connect as $$ transformProducer :: forall i o f m a. MonadRec m => Parallel f m => Producer i m a -> Transformer i o m a -> Producer o m a infixr 2 transformProducer as $~ transformConsumer :: forall i o f m a. MonadRec m => Parallel f m => Transformer i o m a -> Consumer o m a -> Consumer i m a infixr 2 transformConsumer as ~$ composeTransformers :: forall i j k f m a. MonadRec m => Parallel f m => Transformer i j m a -> Transformer j k m a -> Transformer i k m a infixr 2 composeTransformers as ~~
Once composed, the resulting computation can be run using
runFreeT. If you have used
connect to connect a producer and consumer, then you can use
runProcess (a helper function for running a producer/consumer pair via
This example transforms the
nats producer so that instead of producing integers it produces strings, and then connects the resulting producer to the
printer consumer. Once connect, we can use
runProcess to run the pipeline:
main = launchAff $ runProcess ((nats $~ showing) $$ printer)
The producer will emit a value, then yield; this value will satisfy the
await call in the consumer, which will use the value and then yield back to the producer. This process will continue indefinitely.
coroutines documentation is stored in a few places:
- Module documentation is published on Pursuit.
- Written documentation is kept in the docs directory.
- Usage examples can be found in the test suite.
If you get stuck, there are several ways to get help:
- Open an issue if you have encountered a bug or problem.
- Search or start a thread on the PureScript Discourse if you have general questions. You can also ask questions in the
#purescript-beginnerschannels on the Functional Programming Slack (invite link).
You can contribute to
coroutines in several ways:
If you encounter a problem or have a question, please open an issue. We'll do our best to work with you to resolve or answer it.
If you would like to contribute code, tests, or documentation, please read the contributor guide. It's a short, helpful introduction to contributing to this library, including development instructions.
If you have written a library, tutorial, guide, or other resource based on this package, please share it on the PureScript Discourse! Writing libraries and learning resources are a great way to help this library succeed.