📝 Add more examples / uses to the Concurrency section in the documentation guide of Fs2 #3437

Hombre-x · 2024-05-17T20:51:01Z

Helps with #1958

What's included:

Examples for both parEvalMap and parEvalMapUnordered
Reasoning about the three merge variants; it's about when they will halt or not depending on the provided streams
Added a concurrently example using a simple producer-consumer environment (Had to use Scala 2 syntax with implicits, I didn't know how to make mdoc work with Scala 3 at the same time 🤔).
Changed the previous example to make it suitable for parJoin

Added some examples of `parEvalMap` to the *Concurrency* section in the documentation guide.

Added reasoning about `merge` halting variants and an example of `concurrently` using a producer consumer demo (this example only works in Scala 3)

Added a similar consumer-producer example to the `parJoin` method. Also mentioned the `parJoinUnbounded` version.

Apparently, `mdoc` does not have a way to distinguish between Scala 2 and Scala 3 AFAIK. So this commit changes the examples in Scala 3 to Scala 2.

The sample routes in the example are actually linked to the project, so I used the testdata/ folder to place some lorem ipsum files.

The previous one had conflict because of previous definitions in the `concurrently` one. The new example also runs three streams at the same time to differentiate it a bit more from `merge`.

site/guide.md

mpilquist · 2024-05-22T09:47:33Z

site/guide.md

+ Path("sample_file_part3.txt"),
+).map(Path("testdata") / _)
+
+def loadFile(path: Path): IO[String] = 


I don't love this example for a couple of reasons:

concurrency not really adding any benefit given all actions are against the file system

unnecessarily parses the file contents to a string

unnecessarily loads full file contents in memory

if those downsides aren't important, then the same thing could be done entirely with IO and without Stream

Here's a full streaming version of this file concat routine that maintains a streaming nature and avoids parsing:

def concat(inPaths: Stream[IO, Path], outPath: Path): Stream[IO, Nothing] = inPaths.flatMap(p => Files[IO].readAll(p)).through(Files[IO].writeAll(outPath))

You are right, I did It just as an excuse to maintain the order of the original stream 🤔 let me thing of a better example!

Changed tree 🌳 to three in typo on `merge` Co-authored-by: Michael Pilquist <[email protected]>

Changed 'optimizations' word from british to american english Co-authored-by: Michael Pilquist <[email protected]>

Removed unused imports in various examples before `mdoc:reset` Co-authored-by: Michael Pilquist <[email protected]>

Changed the example as suggested. Now it does not parse the bytes, nor reads from the filesystem. Instead it downloads some files from The Project Gutenberg and creates an `InputStream` from each one of them. Co-authored-by: Michael Pilquist <[email protected]>

Hombre-x · 2024-05-28T17:47:45Z

I sometimes forget that GitHub don't usually notifies the commits in PRs, so changes are done @mpilquist!

mpilquist · 2024-05-30T11:47:08Z

Thanks!

Hombre-x added 6 commits May 16, 2024 18:38

📝 Edited guide.md

6a26375

Added some examples of `parEvalMap` to the *Concurrency* section in the documentation guide.

📝 Edited guide.md

43fc6c2

Added reasoning about `merge` halting variants and an example of `concurrently` using a producer consumer demo (this example only works in Scala 3)

📝 Edited guide.md

cb6929a

Added a similar consumer-producer example to the `parJoin` method. Also mentioned the `parJoinUnbounded` version.

🐛📝 Changed Scala 3 code to version 2

43036a3

Apparently, `mdoc` does not have a way to distinguish between Scala 2 and Scala 3 AFAIK. So this commit changes the examples in Scala 3 to Scala 2.

🐛📝 Added test files and changed parEvalMap example

c777ae8

The sample routes in the example are actually linked to the project, so I used the testdata/ folder to place some lorem ipsum files.

🐛📝 Changed the parJoin example

a65f9de

The previous one had conflict because of previous definitions in the `concurrently` one. The new example also runs three streams at the same time to differentiate it a bit more from `merge`.

Hombre-x changed the title ~~📝 Add more examples / uses to the Concurrency section of the documentation guide of Fs2~~ 📝 Add more examples / uses to the Concurrency section in the documentation guide of Fs2 May 18, 2024