OS-Lib

// Make sure working directory exists and is empty
val wd = os.pwd/"out/splash"
os.remove.all(wd)
os.makeDir.all(wd)

// Read/write files
os.write(wd/"file.txt", "hello")
os.read(wd/"file.txt") ==> "hello"

// Perform filesystem operations
os.copy(wd/"file.txt", wd/"copied.txt")
os.list(wd) ==> Seq(wd/"copied.txt", wd/"file.txt")

// Invoke subprocesses
val invoked = os.proc("cat", wd/"file.txt", wd/"copied.txt").call(cwd = wd)
invoked.out.trim ==> "hellohello"

// Chain multiple subprocess' stdin/stdout together
val curl = os.proc("curl", "-L" , "https://git.io/fpvpS").spawn(stderr = os.Inherit)
val gzip = os.proc("gzip", "-n").spawn(stdin = curl.stdout)
val sha = os.proc("shasum", "-a", "256").spawn(stdin = gzip.stdout)
sha.stdout.trim ==> "acc142175fa520a1cb2be5b97cbbe9bea092e8bba3fe2e95afa645615908229e  -"

OS-Lib is a simple Scala interface to common OS filesystem and subprocess APIs. OS-Lib aims to make working with files and processes in Scala as simple as any scripting language, while still providing the safety, flexibility and performance you would expect from Scala.

OS-Lib aims to be a complete replacement for the java.nio.file.Files/java.nio.file.Paths, java.lang.ProcessBuilder scala.io and scala.sys APIs. You should not need to drop down to underlying Java APIs, as OS-Lib exposes all relevant capabilities in an intuitive and performant way. OS-Lib has no dependencies and is unopinionated: it exposes the underlying APIs in a concise but straightforward way, without introducing it’s own idiosyncrasies, quirks, or clever DSLs.

If you use OS-Lib and like it, you will probably enjoy the following book by the Author:

Hands-on Scala Programming

Hands-on Scala has uses OS-Lib extensively throughout the book, and has the entirety of Chapter 7: Files and Subprocesses dedicated to OS-Lib. Hands-on Scala is a great way to level up your skills in Scala in general and OS-Lib in particular.

You can also support it by donating to our Patreon:

https://www.patreon.com/lihaoyi

For a hands-on introduction to the library, take a look at these two blog posts:

Table of Contents

Getting Started
Cookbook
Operations
Spawning Pipelines of Subprocesses
- Watching for Changes
  - os.watch.watch
Data Types
Changelog
- 0.10.7
- 0.10.6
- 0.10.5
- 0.10.4
- 0.10.3
- 0.10.2
- 0.10.1
- 0.10.0
- 0.9.3 - 2024-01-01
- 0.9.2 - 2023-11-05
- 0.9.1 - 2023-03-07
- 0.9.0 - 2022-11-28
- Older releases

Getting Started

To begin using OS-Lib, first add it as a dependency to your project’s build:

// Mill
ivy"com.lihaoyi::os-lib:0.10.7"
// SBT
"com.lihaoyi" %% "os-lib" % "0.10.7"

API Documentation (Scala 3)

Cookbook

Most operation in OS-Lib take place on os.Paths, which are constructed from a base path or working directory wd. Most often, the first thing to do is to define a wd path representing the folder you want to work with:

val wd = os.pwd / "my-test-folder"

You can of course multiple base paths, to use in different parts of your program where convenient, or simply work with one of the pre-defined paths os.pwd, os.root, or os.home.

Concatenate text files

// Find and concatenate all .txt files directly in the working directory
os.write(
  wd / "all.txt",
  os.list(wd).filter(_.ext == "txt").map(os.read)
)

os.read(wd / "all.txt") ==>
  """I am cowI am cow
    |Hear me moo
    |I weigh twice as much as you
    |And I look good on the barbecue""".stripMargin

Spawning a subprocess on multiple files

// Find and concatenate all .txt files directly in the working directory using `cat`
os.proc("cat", os.list(wd).filter(_.ext == "txt")).call(stdout = wd / "all.txt")

os.read(wd / "all.txt") ==>
  """I am cowI am cow
    |Hear me moo
    |I weigh twice as much as you
    |And I look good on the barbecue""".stripMargin

Curl URL to temporary file

// Curl to temporary file
val temp = os.temp()
os.proc("curl", "-L" , "https://git.io/fpfTs").call(stdout = temp)

os.size(temp) ==> 53814

// Curl to temporary file
val temp2 = os.temp()
val proc = os.proc("curl", "-L" , "https://git.io/fpfTJ").spawn()

os.write.over(temp2, proc.stdout)
os.size(temp2) ==> 53814

Recursive line count

// Line-count of all .txt files recursively in wd
val lineCount = os.walk(wd)
  .filter(_.ext == "txt")
  .map(os.read.lines)
  .map(_.size)
  .sum

lineCount ==> 9

Largest Three Files

// Find the largest three files in the given folder tree
val largestThree = os.walk(wd)
  .filter(os.isFile(_, followLinks = false))
  .map(x => os.size(x) -> x).sortBy(-_._1)
  .take(3)

largestThree ==> Seq(
  (711, wd / "misc/binary.png"),
  (81, wd / "Multi Line.txt"),
  (22, wd / "folder1/one.txt")
)

Moving files out of folder

// Move all files inside the "misc" folder out of it
import os./
os.list(wd / "misc").map(os.move.matching { case p/"misc"/x => p/x } )

Calculate word frequencies

// Calculate the word frequency of all the text files in the folder tree
def txt = os.walk(wd).filter(_.ext == "txt").map(os.read)
def freq(s: Seq[String]) = s.groupBy(x => x).mapValues(_.length).toSeq
val map = freq(txt.flatMap(_.split("[^a-zA-Z0-9_]"))).sortBy(-_._2)
map

Operations

Reading & Writing

`os.read`

os.read(arg: os.ReadablePath): String
os.read(arg: os.ReadablePath, charSet: Codec): String
os.read(arg: os.Path,
        offset: Long = 0,
        count: Int = Int.MaxValue,
        charSet: Codec = java.nio.charset.StandardCharsets.UTF_8): String

Reads the contents of a os.Path or other os.Source as a java.lang.String. Defaults to reading the entire file as UTF-8, but you can also select a different charSet to use, and provide an offset/count to read from if the source supports seeking.

os.read(wd / "File.txt") ==> "I am cow"
os.read(wd / "folder1/one.txt") ==> "Contents of folder one"
os.read(wd / "Multi Line.txt") ==>
  """I am cow
    |Hear me moo
    |I weigh twice as much as you
    |And I look good on the barbecue""".stripMargin

`os.read.bytes`

os.read.bytes(arg: os.ReadablePath): Array[Byte]
os.read.bytes(arg: os.Path, offset: Long, count: Int): Array[Byte]

Reads the contents of a os.Path or os.Source as an Array[Byte]; you can provide an offset/count to read from if the source supports seeking.

os.read.bytes(wd / "File.txt") ==> "I am cow".getBytes
os.read.bytes(wd / "misc/binary.png").length ==> 711

`os.read.chunks`

os.read.chunks(p: ReadablePath, chunkSize: Int): os.Generator[(Array[Byte], Int)]
os.read.chunks(p: ReadablePath, buffer: Array[Byte]): os.Generator[(Array[Byte], Int)]

Reads the contents of the given path in chunks of the given size; returns a generator which provides a byte array and an offset into that array which contains the data for that chunk. All chunks will be of the given size, except for the last chunk which may be smaller.

Note that the array returned by the generator is shared between each callback; make sure you copy the bytes/array somewhere else if you want to keep them around.

Optionally takes in a provided input buffer instead of a chunkSize, allowing you to re-use the buffer between invocations.

val chunks = os.read.chunks(wd / "File.txt", chunkSize = 2)
  .map{case (buf, n) => buf.take(n).toSeq } // copy the buffer to save the data
  .toSeq

chunks ==> Seq(
  Seq[Byte]('I', ' '),
  Seq[Byte]('a', 'm'),
  Seq[Byte](' ', 'c'),
  Seq[Byte]('o', 'w')
)

`os.read.lines`

os.read.lines(arg: os.ReadablePath): IndexedSeq[String]
os.read.lines(arg: os.ReadablePath, charSet: Codec): IndexedSeq[String]

Reads the given os.Path or other os.Source as a string and splits it into lines; defaults to reading as UTF-8, which you can override by specifying a charSet.

os.read.lines(wd / "File.txt") ==> Seq("I am cow")
os.read.lines(wd / "Multi Line.txt") ==> Seq(
  "I am cow",
  "Hear me moo",
  "I weigh twice as much as you",
  "And I look good on the barbecue"
)

`os.read.lines.stream`

os.read.lines(arg: os.ReadablePath): os.Generator[String]
os.read.lines(arg: os.ReadablePath, charSet: Codec): os.Generator[String]

Identical to os.read.lines, but streams the results back to you in a os.Generator rather than accumulating them in memory. Useful if the file is large.

os.read.lines.stream(wd / "File.txt").count() ==> 1
os.read.lines.stream(wd / "Multi Line.txt").count() ==> 4

// Streaming the lines to the console
for(line <- os.read.lines.stream(wd / "Multi Line.txt")){
  println(line)
}

`os.read.inputStream`

os.read.inputStream(p: ReadablePath): java.io.InputStream

Opens a java.io.InputStream to read from the given file.

val is = os.read.inputStream(wd / "File.txt") // ==> "I am cow"
is.read() ==> 'I'
is.read() ==> ' '
is.read() ==> 'a'
is.read() ==> 'm'
is.read() ==> ' '
is.read() ==> 'c'
is.read() ==> 'o'
is.read() ==> 'w'
is.read() ==> -1
is.close()

`os.read.stream`

os.read.stream(p: ReadablePath): geny.Readable

Opens a geny.Readable to read from the given file. This allows you to stream data to any other library that supports Readable without buffering the data in memory, e.g. parsing it via FastParse, deserializing it via uPickle, uploading it via Requests-Scala, etc.

val readable: geny.Readable = os.read.stream(wd / "File.json")

requests.post("https://httpbin.org/post", data = readable)

upickle.default.read(readable)

ujson.read(readable)

`os.write`

os.write(target: Path,
         data: os.Source,
         perms: PermSet = null,
         createFolders: Boolean = false): Unit

Writes data from the given file or os.Source to a file at the target os.Path. You can specify the filesystem permissions of the newly created file by passing in a os.PermSet.

This throws an exception if the file already exists. To over-write or append to an existing file, see os.write.over or os.write.append.

By default, this doesn’t create enclosing folders; you can enable this behavior by setting createFolders = true

os.write(wd / "New File.txt", "New File Contents")
os.read(wd / "New File.txt") ==> "New File Contents"

os.write(wd / "NewBinary.bin", Array[Byte](0, 1, 2, 3))
os.read.bytes(wd / "NewBinary.bin") ==> Array[Byte](0, 1, 2, 3)

`os.write.append`

os.write.append(target: Path,
                data: os.Source,
                perms: PermSet = null,
                createFolders: Boolean = false): Unit

Similar to os.write, except if the file already exists this appends the written data to the existing file contents.

os.read(wd / "File.txt") ==> "I am cow"

os.write.append(wd / "File.txt", ", hear me moo")
os.read(wd / "File.txt") ==> "I am cow, hear me moo"

os.write.append(wd / "File.txt", ",\nI weigh twice as much as you")
os.read(wd / "File.txt") ==>
  "I am cow, hear me moo,\nI weigh twice as much as you"

os.read.bytes(wd / "misc/binary.png").length ==> 711
os.write.append(wd / "misc/binary.png", Array[Byte](1, 2, 3))
os.read.bytes(wd / "misc/binary.png").length ==> 714

`os.write.over`

os.write.over(target: Path,
              data: os.Source,
              perms: PermSet = null,
              offset: Long = 0,
              createFolders: Boolean = false,
              truncate: Boolean = true): Unit

Similar to os.write, except if the file already exists this over-writes the existing file contents. You can also pass in truncate = false to avoid truncating the file if the new contents is shorter than the old contents, and an offset to the file you want to write to.

os.read(wd / "File.txt") ==> "I am cow"
os.write.over(wd / "File.txt", "You are cow")

os.read(wd / "File.txt") ==> "You are cow"

os.write.over(wd / "File.txt", "We ", truncate = false)
os.read(wd / "File.txt") ==> "We  are cow"

os.write.over(wd / "File.txt", "s", offset = 8, truncate = false)
os.read(wd / "File.txt") ==> "We  are sow"

`os.write.outputStream`

os.write.outputStream(target: Path,
                      perms: PermSet = null,
                      createFolders: Boolean = false,
                      openOptions: Seq[OpenOption] = Seq(CREATE, WRITE))

Open a java.io.OutputStream to write to the given file.

val out = os.write.outputStream(wd / "New File.txt")
out.write('H')
out.write('e')
out.write('l')
out.write('l')
out.write('o')
out.close()

os.read(wd / "New File.txt") ==> "Hello"

`os.truncate`

os.truncate(p: Path, size: Long): Unit

Truncate the given file to the given size. If the file is smaller than the given size, does nothing.

os.read(wd / "File.txt") ==> "I am cow"

os.truncate(wd / "File.txt", 4)
os.read(wd / "File.txt") ==> "I am"

Listing & Walking

`os.list`

os.list(p: Path): IndexedSeq[Path]
os.list(p: Path, sort: Boolean = true): IndexedSeq[Path]

Returns all the files and folders directly within the given folder. If the given path is not a folder, raises an error. Can be called via os.list.stream to stream the results. To list files recursively, use os.walk.

For convenience os.list sorts the entries in the folder before returning them. You can disable sorted by passing in the flag sort = false.

os.list(wd / "folder1") ==> Seq(wd / "folder1/one.txt")
os.list(wd / "folder2") ==> Seq(
  wd / "folder2/nestedA",
  wd / "folder2/nestedB"
)

`os.list.stream`

os.list.stream(p: Path): os.Generator[Path]

Similar to os.list, except provides a os.Generator of results rather than accumulating all of them in memory. Useful if the result set is large.

os.list.stream(wd / "folder2").count() ==> 2

// Streaming the listed files to the console
for(line <- os.list.stream(wd / "folder2")){
  println(line)
}

`os.walk`

os.walk(path: Path,
        skip: Path => Boolean = _ => false,
        preOrder: Boolean = true,
        followLinks: Boolean = false,
        maxDepth: Int = Int.MaxValue,
        includeTarget: Boolean = false): IndexedSeq[Path]

Recursively walks the given folder and returns the paths of every file or folder within.

You can pass in a skip callback to skip files or folders you are not interested in. This can avoid walking entire parts of the folder hierarchy, saving time as compared to filtering them after the fact.

By default, the paths are returned as a pre-order traversal: the enclosing folder is occurs first before any of it’s contents. You can pass in preOrder = false to turn it into a post-order traversal, such that the enclosing folder occurs last after all it’s contents.

os.walk returns but does not follow symlinks; pass in followLinks = true to override that behavior. You can also specify a maximum depth you wish to walk via the maxDepth parameter.

os.walk does not include the path given to it as part of the traversal by default. Pass in includeTarget = true to make it do so. The path appears at the start of the traversal of preOrder = true, and at the end of the traversal if preOrder = false.

os.walk(wd / "folder1") ==> Seq(wd / "folder1/one.txt")

os.walk(wd / "folder1", includeTarget = true) ==> Seq(
  wd / "folder1",
  wd / "folder1/one.txt"
)

os.walk(wd / "folder2") ==> Seq(
  wd / "folder2/nestedA",
  wd / "folder2/nestedA/a.txt",
  wd / "folder2/nestedB",
  wd / "folder2/nestedB/b.txt"
)

os.walk(wd / "folder2", preOrder = false) ==> Seq(
  wd / "folder2/nestedA/a.txt",
  wd / "folder2/nestedA",
  wd / "folder2/nestedB/b.txt",
  wd / "folder2/nestedB"
)

os.walk(wd / "folder2", maxDepth = 1) ==> Seq(
  wd / "folder2/nestedA",
  wd / "folder2/nestedB"
)

os.walk(wd / "folder2", skip = _.last == "nestedA") ==> Seq(
  wd / "folder2/nestedB",
  wd / "folder2/nestedB/b.txt"
)

`os.walk.attrs`

os.walk.attrs(path: Path,
              skip: (Path, os.StatInfo) => Boolean = (_, _) => false,
              preOrder: Boolean = true,
              followLinks: Boolean = false,
              maxDepth: Int = Int.MaxValue,
              includeTarget: Boolean = false): IndexedSeq[(Path, os.StatInfo)]

Similar to os.walk, except it also provides the os.StatInfo filesystem metadata of every path that it returns. Can save time by allowing you to avoid querying the filesystem for metadata later. Note that os.StatInfo does not include filesystem ownership and permissions data; use os.stat.posix on the path if you need those attributes.

val filesSortedBySize = os.walk.attrs(wd / "misc", followLinks = true)
  .sortBy{case (p, attrs) => attrs.size}
  .collect{case (p, attrs) if attrsisFile => p}

filesSortedBySize ==> Seq(
  wd / "misc/echo",
  wd / "misc/file-symlink",
  wd / "misc/echo_with_wd",
  wd / "misc/folder-symlink/one.txt",
  wd / "misc/binary.png"
)

`os.walk.stream`

os.walk.stream(path: Path,
              skip: Path => Boolean = _ => false,
              preOrder: Boolean = true,
              followLinks: Boolean = false,
              maxDepth: Int = Int.MaxValue,
              includeTarget: Boolean = false): os.Generator[Path]

Similar to os.walk, except returns a os.Generator of the results rather than accumulating them in memory. Useful if you are walking very large folder hierarchies, or if you wish to begin processing the output even before the walk has completed.

os.walk.stream(wd / "folder1").count() ==> 1

os.walk.stream(wd / "folder2").count() ==> 4

os.walk.stream(wd / "folder2", skip = _.last == "nestedA").count() ==> 2

`os.walk.stream.attrs`

os.walk.stream.attrs(path: Path,
                     skip: (Path, os.StatInfo) => Boolean = (_, _) => false,
                     preOrder: Boolean = true,
                     followLinks: Boolean = false,
                     maxDepth: Int = Int.MaxValue,
                     includeTarget: Boolean = false): os.Generator[(Path, os.StatInfo)]

Similar to os.walk.stream, except it also provides the filesystem metadata of every path that it returns. Can save time by allowing you to avoid querying the filesystem for metadata later.

def totalFileSizes(p: os.Path) = os.walk.stream.attrs(p)
  .collect{case (p, attrs) if attrs.isFile => attrs.size}
  .sum

totalFileSizes(wd / "folder1") ==> 22
totalFileSizes(wd / "folder2") ==> 40

Manipulating Files & Folders

`os.exists`

os.exists(p: Path, followLinks: Boolean = true): Boolean

Checks if a file or folder exists at the specified path

os.exists(wd / "File.txt") ==> true
os.exists(wd / "folder1") ==> true
os.exists(wd / "doesnt-exist") ==> false

os.exists(wd / "misc/file-symlink") ==> true
os.exists(wd / "misc/folder-symlink") ==> true
os.exists(wd / "misc/broken-symlink") ==> false
os.exists(wd / "misc/broken-symlink", followLinks = false) ==> true

`os.move`

os.move(from: Path, to: Path): Unit
os.move(from: Path, to: Path, createFolders: Boolean): Unit

Moves a file or folder from one path to another. Errors out if the destination path already exists, or is within the source path.

os.list(wd / "folder1") ==> Seq(wd / "folder1/one.txt")
os.move(wd / "folder1/one.txt", wd / "folder1/first.txt")
os.list(wd / "folder1") ==> Seq(wd / "folder1/first.txt")

os.list(wd / "folder2") ==> Seq(wd / "folder2/nestedA", wd / "folder2/nestedB")
os.move(wd / "folder2/nestedA", wd / "folder2/nestedC")
os.list(wd / "folder2") ==> Seq(wd / "folder2/nestedB", wd / "folder2/nestedC")

os.read(wd / "File.txt") ==> "I am cow"
os.move(wd / "Multi Line.txt", wd / "File.txt", replaceExisting = true)
os.read(wd / "File.txt") ==>
  """I am cow
    |Hear me moo
    |I weigh twice as much as you
    |And I look good on the barbecue""".stripMargin

`os.move.matching`

os.move.matching(t: PartialFunction[Path, Path]): PartialFunction[Path, Unit]

os.move can also be used as a transformer, via os.move.matching. This lets you use .map or .collect on a list of paths, and move all of them at once, e.g. to rename all .txt files within a folder tree to .data:

import os.{GlobSyntax, /}
os.walk(wd / "folder2") ==> Seq(
  wd / "folder2/nestedA",
  wd / "folder2/nestedA/a.txt",
  wd / "folder2/nestedB",
  wd / "folder2/nestedB/b.txt"
)

os.walk(wd/'folder2).collect(os.move.matching{case p/g"$x.txt" => p/g"$x.data"})

os.walk(wd / "folder2") ==> Seq(
  wd / "folder2/nestedA",
  wd / "folder2/nestedA/a.data",
  wd / "folder2/nestedB",
  wd / "folder2/nestedB/b.data"
)

`os.move.into`

os.move.into(from: Path, to: Path): Unit

Move the given file or folder into the destination folder

os.list(wd / "folder1") ==> Seq(wd / "folder1/one.txt")
os.move.into(wd / "File.txt", wd / "folder1")
os.list(wd / "folder1") ==> Seq(wd / "folder1/File.txt", wd / "folder1/one.txt")

`os.move.over`

os.move.over(from: Path, to: Path): Unit

Move a file or folder from one path to another, and overwrite any file or folder than may already be present at that path

os.list(wd / "folder2") ==> Seq(wd / "folder2/nestedA", wd / "folder2/nestedB")
os.move.over(wd / "folder1", wd / "folder2")
os.list(wd / "folder2") ==> Seq(wd / "folder2/one.txt")

`os.copy`

os.copy(from: Path, to: Path): Unit
os.copy(from: Path, to: Path, createFolders: Boolean): Unit

Copy a file or folder from one path to another. Recursively copies folders with all their contents. Errors out if the destination path already exists, or is within the source path.

os.list(wd / "folder1") ==> Seq(wd / "folder1/one.txt")
os.copy(wd / "folder1/one.txt", wd / "folder1/first.txt")
os.list(wd / "folder1") ==> Seq(wd / "folder1/first.txt", wd / "folder1/one.txt")

os.list(wd / "folder2") ==> Seq(wd / "folder2/nestedA", wd / "folder2/nestedB")
os.copy(wd / "folder2/nestedA", wd / "folder2/nestedC")
os.list(wd / "folder2") ==> Seq(
  wd / "folder2/nestedA",
  wd / "folder2/nestedB",
  wd / "folder2/nestedC"
)

os.read(wd / "File.txt") ==> "I am cow"
os.copy(wd / "Multi Line.txt", wd / "File.txt", replaceExisting = true)
os.read(wd / "File.txt") ==>
  """I am cow
    |Hear me moo
    |I weigh twice as much as you
    |And I look good on the barbecue""".stripMargin
    ```

`os.copy` can also be used as a transformer:

```scala
os.copy.matching(t: PartialFunction[Path, Path]): PartialFunction[Path, Unit]

This lets you use .map or .collect on a list of paths, and copy all of them at once:

paths.map(os.copy.matching{case p/"scala"/file => p/"java"/file})

`os.copy.into`

os.copy.into(from: Path, to: Path): Unit

Copy the given file or folder into the destination folder

os.list(wd / "folder1") ==> Seq(wd / "folder1/one.txt")
os.copy.into(wd / "File.txt", wd / "folder1")
os.list(wd / "folder1") ==> Seq(wd / "folder1/File.txt", wd / "folder1/one.txt")

`os.copy.over`

os.copy.over(from: Path, to: Path): Unit

Similar to os.copy, but if the destination file already exists then overwrite it instead of erroring out.

os.list(wd / "folder2") ==> Seq(wd / "folder2/nestedA", wd / "folder2/nestedB")
os.copy.over(wd / "folder1", wd / "folder2")
os.list(wd / "folder2") ==> Seq(wd / "folder2/one.txt")

`os.copy` with `mergeFolders`

Since 0.7.5

If you want to copy a directory over another but don’t want to overwrite the whole destination directory (and loose it’s content), you can use the mergeFolders option of os.copy.

os.list(wd / "folder1") ==> Seq(wd / "folder1/one.txt")
os.list(wd / "folder2") ==> Seq(wd / "folder2/nestedA", wd / "folder2/nestedB")
os.copy(wd / "folder1", wd / "folder2", mergeFolders = true)
os.list(wd / "folder2") ==> Seq(wd / "folder2/one.txt", wd / "folder2/nestedA", wd / "folder2/nestedB")

`os.makeDir`

os.makeDir(path: Path): Unit
os.makeDir(path: Path, perms: PermSet): Unit

Create a single directory at the specified path. Optionally takes in a os.PermSet to specify the filesystem permissions of the created directory.

Errors out if the directory already exists, or if the parent directory of the specified path does not exist. To automatically create enclosing directories and ignore the destination if it already exists, using os.makeDir.all

os.exists(wd / "new_folder") ==> false
os.makeDir(wd / "new_folder")
os.exists(wd / "new_folder") ==> true

`os.makeDir.all`

os.makeDir.all(path: Path): Unit
os.makeDir.all(path: Path,
               perms: PermSet = null,
               acceptLinkedDirectory: Boolean = true): Unit

Similar to os.makeDir, but automatically creates any necessary enclosing directories if they do not exist, and does not raise an error if the destination path already contains a directory. Also does not raise an error if the destination path contains a symlink to a directory, though you can force it to error out in that case by passing in acceptLinkedDirectory = false

os.exists(wd / "new_folder") ==> false
os.makeDir.all(wd / "new_folder/inner/deep")
os.exists(wd / "new_folder/inner/deep") ==> true

`os.remove`

os.remove(target: Path): Boolean
os.remove(target: Path, checkExists: Boolean = false): Boolean

Remove the target file or folder. Folders need to be empty to be removed; if you want to remove a folder tree recursively, use os.remove.all. Returns true if the file was present before. It will fail with an exception when the file is missing but checkExists is true, or when the directory to remove is not empty.

os.exists(wd / "File.txt") ==> true
os.remove(wd / "File.txt")
os.exists(wd / "File.txt") ==> false

os.exists(wd / "folder1/one.txt") ==> true
os.remove(wd / "folder1/one.txt")
os.remove(wd / "folder1")
os.exists(wd / "folder1/one.txt") ==> false
os.exists(wd / "folder1") ==> false

When removing symbolic links, it is the link that gets removed, and not its destination:

os.remove(wd / "misc/file-symlink")
os.exists(wd / "misc/file-symlink", followLinks = false) ==> false
os.exists(wd / "File.txt", followLinks = false) ==> true

os.remove(wd / "misc/folder-symlink")
os.exists(wd / "misc/folder-symlink", followLinks = false) ==> false
os.exists(wd / "folder1", followLinks = false) ==> true
os.exists(wd / "folder1/one.txt", followLinks = false) ==> true

os.remove(wd / "misc/broken-symlink")
os.exists(wd / "misc/broken-symlink", followLinks = false) ==> false

If you wish to remove the destination of a symlink, use os.readLink.

`os.remove.all`

os.remove.all(target: Path): Unit

Remove the target file or folder; if it is a folder and not empty, recursively removing all it’s contents before deleting it.

os.exists(wd / "folder1/one.txt") ==> true
os.remove.all(wd / "folder1")
os.exists(wd / "folder1/one.txt") ==> false
os.exists(wd / "folder1") ==> false

When removing symbolic links, it is the links that gets removed, and not it’s destination:

os.remove.all(wd / "misc/file-symlink")
os.exists(wd / "misc/file-symlink", followLinks = false) ==> false
os.exists(wd / "File.txt", followLinks = false) ==> true

os.remove.all(wd / "misc/folder-symlink")
os.exists(wd / "misc/folder-symlink", followLinks = false) ==> false
os.exists(wd / "folder1", followLinks = false) ==> true
os.exists(wd / "folder1/one.txt", followLinks = false) ==> true

os.remove.all(wd / "misc/broken-symlink")
os.exists(wd / "misc/broken-symlink", followLinks = false) ==> false

If you wish to remove the destination of a symlink, use os.readLink.

`os.hardlink`

os.hardlink(src: Path, dest: Path, perms): Unit

Create a hardlink to the source path from the destination path

os.hardlink(wd / "File.txt", wd / "Linked.txt")
os.exists(wd / "Linked.txt")
os.read(wd / "Linked.txt") ==> "I am cow"
os.isLink(wd / "Linked.txt") ==> false

`os.symlink`

os.symlink(link: Path, dest: FilePath, perms: PermSet = null): Unit

Create a symbolic to the source path from the destination path. Optionally takes a os.PermSet to customize the filesystem permissions of the symbolic link.

os.symlink(wd / "File.txt", wd / "Linked.txt")
os.exists(wd / "Linked.txt")
os.read(wd / "Linked.txt") ==> "I am cow"
os.isLink(wd / "Linked.txt") ==> true

You can create symlinks with either absolute os.Paths or relative os.RelPaths:

os.symlink(wd / "File.txt", os.rel/ "Linked2.txt")
os.exists(wd / "Linked2.txt")
os.read(wd / "Linked2.txt") ==> "I am cow"
os.isLink(wd / "Linked2.txt") ==> true

Creating absolute and relative symlinks respectively. Relative symlinks are resolved relative to the enclosing folder of the link.

`os.readLink`

os.readLink(src: Path): os.FilePath
os.readLink.absolute(src: Path): os.Path

Returns the immediate destination of the given symbolic link.

os.readLink(wd / "misc/file-symlink") ==> os.up / "File.txt"
os.readLink(wd / "misc/folder-symlink") ==> os.up / "folder1"
os.readLink(wd / "misc/broken-symlink") ==> os.rel / "broken"
os.readLink(wd / "misc/broken-abs-symlink") ==> os.root / "doesnt/exist"

Note that symbolic links can be either absolute os.Paths or relative os.RelPaths, represented by os.FilePath. You can also use os.readLink.absolute to automatically resolve relative symbolic links to their absolute destination:

os.readLink.absolute(wd / "misc/file-symlink") ==> wd / "File.txt"
os.readLink.absolute(wd / "misc/folder-symlink") ==> wd / "folder1"
os.readLink.absolute(wd / "misc/broken-symlink") ==> wd / "misc/broken"
os.readLink.absolute(wd / "misc/broken-abs-symlink") ==> os.root / "doesnt/exist"

`os.followLink`

os.followLink(src: Path): Option[Path]

Attempts to any deference symbolic links in the given path, recursively, and return the canonical path. Returns None if the path cannot be resolved (i.e. some symbolic link in the given path is broken)

os.followLink(wd / "misc/file-symlink") ==> Some(wd / "File.txt")
os.followLink(wd / "misc/folder-symlink") ==> Some(wd / "folder1")
os.followLink(wd / "misc/broken-symlink") ==> None

`os.temp`

os.temp(contents: os.Source = null,
        dir: Path = null,
        prefix: String = null,
        suffix: String = null,
        deleteOnExit: Boolean = true,
        perms: PermSet = null): Path

Creates a temporary file. You can optionally provide a dir to specify where this file lives, file-prefix and file-suffix to customize what it looks like, and a os.PermSet to customize its filesystem permissions.

Passing in a os.Source will initialize the contents of that file to the provided data; otherwise it is created empty.

By default, temporary files are deleted on JVM exit. You can disable that behavior by setting deleteOnExit = false

val tempOne = os.temp("default content")
os.read(tempOne) ==> "default content"
os.write.over(tempOne, "Hello")
os.read(tempOne) ==> "Hello"

`os.temp.dir`

os.temp.dir(dir: Path = null,
            prefix: String = null,
            deleteOnExit: Boolean = true,
            perms: PermSet = null): Path

Creates a temporary directory. You can optionally provide a dir to specify where this file lives, a prefix to customize what it looks like, and a os.PermSet to customize its filesystem permissions.

By default, temporary directories are deleted on JVM exit. You can disable that behavior by setting deleteOnExit = false

val tempDir = os.temp.dir()
os.list(tempDir) ==> Nil
os.write(tempDir / "file", "Hello")
os.list(tempDir) ==> Seq(tempDir / "file")

Filesystem Metadata

`os.stat`

os.stat(p: os.Path, followLinks: Boolean = true): os.StatInfo

Reads in the basic filesystem metadata for the given file. By default, follows symbolic links to read the metadata of whatever the link is pointing at; set followLinks = false to disable that and instead read the metadata of the symbolic link itself.

os.stat(wd / "File.txt").size ==> 8
os.stat(wd / "Multi Line.txt").size ==> 81
os.stat(wd / "folder1").fileType ==> os.FileType.Dir

`os.stat.posix`

os.stat.posix(p: os.Path, followLinks: Boolean = true): os.PosixStatInfo

Reads in the posix filesystem metadata for the given file, providing information on permissions and ownership. By default, follows symbolic links to read the metadata of whatever the link is pointing at; set followLinks = false to disable that and instead read the metadata of the symbolic link itself.

`os.isFile`

os.isFile(p: Path, followLinks: Boolean = true): Boolean

Returns true if the given path is a file. Follows symbolic links by default, pass in followLinks = false to not do so.

os.isFile(wd / "File.txt") ==> true
os.isFile(wd / "folder1") ==> false

os.isFile(wd / "misc/file-symlink") ==> true
os.isFile(wd / "misc/folder-symlink") ==> false
os.isFile(wd / "misc/file-symlink", followLinks = false) ==> false

`os.isDir`

os.isDir(p: Path, followLinks: Boolean = true): Boolean

Returns true if the given path is a folder. Follows symbolic links by default, pass in followLinks = false to not do so.

os.isDir(wd / "File.txt") ==> false
os.isDir(wd / "folder1") ==> true

os.isDir(wd / "misc/file-symlink") ==> false
os.isDir(wd / "misc/folder-symlink") ==> true
os.isDir(wd / "misc/folder-symlink", followLinks = false) ==> false

`os.isLink`

os.isLink(p: Path, followLinks: Boolean = true): Boolean

Returns true if the given path is a symbolic link. Follows symbolic links by default, pass in followLinks = false to not do so.

os.isLink(wd / "misc/file-symlink") ==> true
os.isLink(wd / "misc/folder-symlink") ==> true
os.isLink(wd / "folder1") ==> false

`os.size`

os.size(p: Path): Long

Returns the size of the given file, in bytes

os.size(wd / "File.txt") ==> 8
os.size(wd / "Multi Line.txt") ==> 81

`os.mtime`

os.mtime(p: Path): Long
os.mtime.set(p: Path, millis: Long): Unit

Gets or sets the last-modified timestamp of the given file, in milliseconds

os.mtime.set(wd / "File.txt", 0)
os.mtime(wd / "File.txt") ==> 0

os.mtime.set(wd / "File.txt", 90000)
os.mtime(wd / "File.txt") ==> 90000
os.mtime(wd / "misc/file-symlink") ==> 90000

os.mtime.set(wd / "misc/file-symlink", 70000)
os.mtime(wd / "File.txt") ==> 70000
os.mtime(wd / "misc/file-symlink") ==> 70000
assert(os.mtime(wd / "misc/file-symlink", followLinks = false) != 40000)

Filesystem Permissions

`os.perms`

os.perms(p: Path, followLinks: Boolean = true): PermSet
os.perms.set(p: Path, arg2: PermSet): Unit

Gets or sets the filesystem permissions of the given file or folder, as an os.PermSet.

Note that if you want to create a file or folder with a given set of permissions, you can pass in an os.PermSet to os.write or os.makeDir. That will ensure the file or folder is created atomically with the given permissions, rather than being created with the default set of permissions and having os.perms.set over-write them later

os.perms.set(wd / "File.txt", "rwxrwxrwx")
os.perms(wd / "File.txt").toString() ==> "rwxrwxrwx"
os.perms(wd / "File.txt").toInt() ==> Integer.parseInt("777", 8)

os.perms.set(wd / "File.txt", Integer.parseInt("755", 8))
os.perms(wd / "File.txt").toString() ==> "rwxr-xr-x"

os.perms.set(wd / "File.txt", "r-xr-xr-x")
os.perms.set(wd / "File.txt", Integer.parseInt("555", 8))

`os.owner`

os.owner(p: Path, followLinks: Boolean = true): UserPrincipal
os.owner.set(arg1: Path, arg2: UserPrincipal): Unit
os.owner.set(arg1: Path, arg2: String): Unit

Gets or sets the owner of the given file or folder. Note that your process needs to be running as the root user in order to do this.

val originalOwner = os.owner(wd / "File.txt")

os.owner.set(wd / "File.txt", "nobody")
os.owner(wd / "File.txt").getName ==> "nobody"

os.owner.set(wd / "File.txt", originalOwner)

`os.group`

os.group(p: Path, followLinks: Boolean = true): GroupPrincipal
os.group.set(arg1: Path, arg2: GroupPrincipal): Unit
os.group.set(arg1: Path, arg2: String): Unit

Gets or sets the owning group of the given file or folder. Note that your process needs to be running as the root user in order to do this.

val originalOwner = os.owner(wd / "File.txt")

os.owner.set(wd / "File.txt", "nobody")
os.owner(wd / "File.txt").getName ==> "nobody"

os.owner.set(wd / "File.txt", originalOwner)

Spawning Subprocesses

Subprocess are spawned using os.call(cmd: os.Shellable, ...) or os.spawn(cmd: os.Shellable, ...) calls, where the cmd: Shellable sets up the basic command you wish to run and .foo(...) specifies how you want to run it. os.Shellable represents a value that can make up part of your subprocess command, and the following values can be used as os.Shellables:

java.lang.String
scala.Symbol
os.Path
os.RelPath
T: Numeric
Iterable[T]s of any of the above
TupleN[T1, T2, …Tn]s of any of the above

Most of the subprocess commands also let you redirect the subprocess’s stdin/stdout/stderr streams via os.ProcessInput or os.ProcessOutput values: whether to inherit them from the parent process, stream them into buffers, or output them to files. The following values are common to both input and output:

os.Pipe: the default, this connects the subprocess’s stream to the parent process via pipes; if used on its stdin this lets the parent process write to the subprocess via os.SubProcess#stdin, and if used on its stdout it lets the parent process read from the subprocess via os.SubProcess#stdout and os.SubProcess#stderr.
os.Inherit: inherits the stream from the parent process. This lets the subprocess read directly from the parent process’s standard input or write directly to the parent process’s standard output or error. os.Inherit can be redirected on a threadlocal basis via os.Inherit.in, .out, or .err.
os.InheritRaw: identical to os.Inherit, but without being affected by redirects.
os.Path: connects the subprocess’s stream to the given filesystem path, reading its standard input from a file or writing its standard output/error to the file.

In addition, you can pass any os.Sources to a Subprocess’s stdin (Strings, InputStreams, Array[Byte]s, …), and pass in a os.ProcessOutput value to stdout/stderr to register callbacks that are run when output is received on those streams.

Often, if you are only interested in capturing the standard output of the subprocess but want any errors sent to the console, you might set stderr = os.Inherit while leaving stdout = os.Pipe.

`os.call`

os.call(cmd: os.Shellable,
        cwd: Path = null,
        env: Map[String, String] = null,
        stdin: ProcessInput = Pipe,
        stdout: ProcessOutput = Pipe,
        stderr: ProcessOutput = Pipe,
        mergeErrIntoOut: Boolean = false,
        timeout: Long = Long.MaxValue,
        check: Boolean = true,
        propagateEnv: Boolean = true): os.CommandResult

Also callable via `os.proc(cmd).call(…)`

Invokes the given subprocess like a function, passing in input and returning a CommandResult. You can then call result.exitCode to see how it exited, or result.out.bytes or result.err.string to access the aggregated stdout and stderr of the subprocess in a number of convenient ways.

call provides a number of parameters that let you configure how the subprocess is run:

cwd: the working directory of the subprocess
env: any additional environment variables you wish to set in the subprocess
stdin: any data you wish to pass to the subprocess’s standard input
stdout/stderr: these are os.Redirects that let you configure how the processes output/error streams are configured.
mergeErrIntoOut: merges the subprocess’s stderr stream into it’s stdout
timeout: how long to wait for the subprocess to complete
check: disable this to avoid throwing an exception if the subprocess fails with a non-zero exit code
propagateEnv: disable this to avoid passing in this parent process’s environment variables to the subprocess

Note that redirecting stdout/stderr elsewhere means that the respective CommandResult#out/CommandResult#err values will be empty.

val res = os.call(cmd = ('ls, wd/"folder2"))

res.exitCode ==> 0

res.out.text() ==>
  """nestedA
    |nestedB
    |""".stripMargin

res.out.trim() ==>
  """nestedA
    |nestedB""".stripMargin

res.out.lines() ==> Seq(
  "nestedA",
  "nestedB"
)

res.out.bytes


// Non-zero exit codes throw an exception by default
val thrown = intercept[os.SubprocessException]{
  os.call(cmd = ('ls, "doesnt-exist"), cwd = wd)
}

assert(thrown.result.exitCode != 0)

// Though you can avoid throwing by setting `check = false`
val fail = os.call(cmd = ('ls, "doesnt-exist"), cwd = wd, check = false)

assert(fail.exitCode != 0)


fail.out.text() ==> ""

assert(fail.err.text().contains("No such file or directory"))

// You can pass in data to a subprocess' stdin
val hash = os.call(cmd = ("shasum", "-a", "256"), stdin = "Hello World")
hash.out.trim() ==> "a591a6d40bf420404a011733cfb7b190d62c65bf0bcda32b57b277d9ad9f146e  -"

// Taking input from a file and directing output to another file
os.call(cmd = ("base64"), stdin = wd / "File.txt", stdout = wd / "File.txt.b64")

os.read(wd / "File.txt.b64") ==> "SSBhbSBjb3c="

If you want to spawn an interactive subprocess, such as vim, less, or a python shell, set all of stdin/stdout/stderr to os.Inherit:

os.proc("vim").call(stdin = os.Inherit, stdout = os.Inherit, stderr = os.Inherit)

Note that by customizing stdout and stderr, you can use the results of os.proc.call in a streaming fashion, either on groups of bytes:

var lineCount = 1
os.call(
  cmd = ('find, "."),
  cwd = wd,
  stdout = os.ProcessOutput(
    (buf, len) => lineCount += buf.slice(0, len).count(_ == '\n')
  ),
)

Or on lines of output:

lineCount ==> 22
var lineCount = 1
os.call(
  cmd = ('find, "."),
  cwd = wd,
  stdout = os.ProcessOutput.Readlines(
    line => lineCount += 1
  ),
)
lineCount ==> 22

`os.spawn`

os.spawn(cmd: os.Shellable,
         cwd: Path = null,
         env: Map[String, String] = null,
         stdin: os.ProcessInput = os.Pipe,
         stdout: os.ProcessOutput = os.Pipe,
         stderr: os.ProcessOutput = os.Pipe,
         mergeErrIntoOut: Boolean = false,
         propagateEnv: Boolean = true): os.SubProcess

Also callable via `os.proc(cmd).spawn(…)`

The most flexible of the os.proc calls, os.spawn simply configures and starts a subprocess, and returns it as a os.SubProcess. os.SubProcess is a simple wrapper around java.lang.Process, which provides stdin, stdout, and stderr streams for you to interact with however you like. e.g. You can sending commands to it’s stdin and reading from it’s stdout.

To implement pipes, you can spawn a process, take its stdout, and pass it as the stdin of a second spawned process.

Note that if you provide ProcessOutput callbacks to stdout/stderr, the calls to those callbacks take place on newly spawned threads that execute in parallel with the main thread. Thus make sure any data processing you do in those callbacks is thread safe!

stdin, stdout and stderr are java.lang.OutputStreams and java.lang.InputStreams enhanced with the .writeLine(s: String)/.readLine() methods for easy reading and writing of character and line-based data.

// Start a long-lived python process which you can communicate with
val sub = os.spawn(
  cmd = ("python", "-u", "-c", "while True: print(eval(raw_input()))"),
  cwd = wd
)

// Sending some text to the subprocess
sub.stdin.write("1 + 2")
sub.stdin.writeLine("+ 4")
sub.stdin.flush()
sub.stdout.readLine() ==> "7"

sub.stdin.write("'1' + '2'")
sub.stdin.writeLine("+ '4'")
sub.stdin.flush()
sub.stdout.readLine() ==> "124"

// Sending some bytes to the subprocess
sub.stdin.write("1 * 2".getBytes)
sub.stdin.write("* 4\n".getBytes)
sub.stdin.flush()
sub.stdout.read() ==> '8'.toByte

sub.destroy()

// You can chain multiple subprocess' stdin/stdout together
val curl = os.spawn(cmd = ("curl", "-L" , "https://git.io/fpfTs"), stderr = os.Inherit)
val gzip = os.spawn(cmd = ("gzip", "-n"), stdin = curl.stdout)
val sha = os.spawn(cmd = ("shasum", "-a", "256"), stdin = gzip.stdout)
sha.stdout.trim ==> "acc142175fa520a1cb2be5b97cbbe9bea092e8bba3fe2e95afa645615908229e  -"

Customizing the default environment

Client-server CLI applications sometimes want to run subprocesses on the server based on the environment of the client. It is possible to customize the default environment passed to subprocesses by setting the os.SubProcess.env threadlocal:

val clientEnvironment: Map[String, String] = ???
os.SubProcess.env.withValue(clientEnvironment) {
  os.call(command) // clientEnvironment is passed by default instead of the system environment
}

Spawning Pipelines of Subprocesses

After constructing a subprocess with os.proc, you can use the pipeTo method to pipe its output to another subprocess:

val wc = os.proc("ls", "-l")
  .pipeTo(os.proc("wc", "-l"))
  .call()
  .out.text()

This is equivalent to the shell command ls -l | wc -l. You can chain together as many subprocesses as you like. Note that by using this API you can utilize the broken pipe behaviour of Unix systems. For example, you can take 10 first elements of output from the yes command, and after the head command terminates, the yes command will be terminated as well:

val yes10 = os.proc("yes")
  .pipeTo(os.proc("head", "-n", "10"))
  .call()
  .out.text()

This feature is implemented inside the library and will terminate any process reading the stdin of other process in pipeline on every IO error. This behavior can be disabled via the handleBrokenPipe flag on call and spawn methods. Note that Windows does not support broken pipe behaviour, so a command like`yes` would run forever. handleBrokenPipe is set to false by default on Windows.

Both call and spawn correspond in their behavior to their counterparts in the os.proc, but spawn returns the os.ProcessPipeline instance instead. It offers the same API as SubProcess, but will operate on the set of processes instead of a single one.

Pipefail is enabled by default, so if any of the processes in the pipeline fails, the whole pipeline will have a non-zero exit code. This behavior can be disabled via the pipefail flag on call and spawn methods. Note that the pipefail does not kill the processes in the pipeline, it just sets the exit code of the pipeline to the exit code of the failed process.

Watching for Changes

`os.watch.watch`

os.watch.watch(roots: Seq[os.Path], onEvent: Set[os.Path] => Unit): Unit

// Mill
ivy"com.lihaoyi::os-lib-watch:0.10.7"
// SBT
"com.lihaoyi" %% "os-lib-watch" % "0.10.7"

Efficiently watches the given roots folders for changes. Any time the filesystem is modified within those folders, the onEvent callback is called with the paths to the changed files or folders. Note that os.watch.watch is under a different artifact than the rest of the os.* functions, and you need to add a separate dependency to os-lib-watch in order to pull it in.

Once the call to watch returns, onEvent is guaranteed to receive a an event containing the path for:

Every file or folder that gets created, deleted, updated or moved within the watched folders
For copied or moved folders, the path of the new folder as well as every file or folder within it.
For deleted or moved folders, the root folder which was deleted/moved, but without the paths of every file that was within it at the original location

Note that watch does not provide any additional information about the changes happening within the watched roots folder, apart from the path at which the change happened. It is up to the onEvent handler to query the filesystem and figure out what happened, and what it wants to do.

Here is an example of use from the Ammonite REPL:

@ import $ivy.`com.lihaoyi::os-lib-watch:0.10.7`

@ os.watch.watch(Seq(os.pwd / "out"), paths => println("paths changed: " + paths.mkString(", ")))

@ os.write(os.pwd / "out/i am", "cow")

paths changed: /Users/lihaoyi/Github/Ammonite/out/i am

@ os.move(os.pwd / "out/i am", os.pwd / "out/hear me")

paths changed: /Users/lihaoyi/Github/Ammonite/out/i am,/Users/lihaoyi/Github/Ammonite/out/hear me

@ os.remove.all(os.pwd / "out/version")

paths changed: /Users/lihaoyi/Github/Ammonite/out/version/log,/Users/lihaoyi/Github/Ammonite/out/version/meta.json,/Users/lihaoyi/Github/Ammonite/out/version

Data Types

`os.Path`

OS-Lib uses strongly-typed data-structures to represent filesystem paths. The two basic versions are:

os.Path: an absolute path, starting from the root
os.RelPath: a relative path, not rooted anywhere
os.SubPath: a sub path, without any .. segments, not rooted anywhere

Generally, almost all commands take absolute os.Paths. These are basically java.nio.file.Paths with additional guarantees:

os.Paths are always absolute. Relative paths are a separate type os.RelPath
os.Paths are always canonical. You will never find . or .. segments in them, and never need to worry about calling .normalize before operations.

Absolute paths can be created in a few ways:

// Get the process' Current Working Directory. As a convention
// the directory that "this" code cares about (which may differ
// from the pwd) is called `wd`
val wd = os.pwd

// A path nested inside `wd` in multiple segments
wd / "folder" / "file"

// The RHS of `/` can have multiple segments if-and-only-if it is a literal string
wd / "folder/file"

// A path starting from the root
os.root / "folder/file"

// A path with spaces or other special characters
wd / "My Folder/My File.txt"

// Up one level from the wd
wd / os.up

// Up two levels from the wd
wd / os.up / os.up

When constructing os.Path`s, the right-hand-side of the `/ operator must be either a non-literal a string expression containing a single path segment or a literal string containing one-or-more path segments. If a non-literal string expression on the RHS contains multiple segments, you need to wrap the RHS in an explicit os.RelPath(…) or os.SubPath(…) constructor to tell OS-Lib how to interpret it. The single-segment limitation is intended to avoid the developer accidentally introducing Directory Traversal Attacks or other related bugs when naively constructing paths out of dynamic and potentially untrusted inputs, which is not an issue for literal string since the string value is directly written in the source code and immediately visible.

os.pwd can be modified in certain scopes via the os.dynamicPwd dynamic variable, but best practice is not to change it. Instead simply define a new path, e.g.

val target = os.pwd / "target"

Should be sufficient for most needs.

Above, we made use of the os.pwd built-in path. There are a number of Paths built into OS-Lib:

os.pwd: The current working directory of the process. This can’t be changed in Java, so if you need another path to work with the convention is to define a wd variable.
os.root: The root of the filesystem.
os.home: The home directory of the current user.
os.temp()/os.temp.dir(): Creates a temporary file/folder and returns the path.

`os.RelPath`

os.RelPaths represent relative paths. These are basically defined as:

class RelPath private[ops] (segments0: Array[String], val ups: Int)

The same data structure as Paths, except that they can represent a number of ups before the relative path is applied. They can be created in the following ways:

// The path "folder/file" in multiple segments
val rel1 = os.rel / "folder" / "file"
// RHS of `/` can have multiple segments if-and-only-if it is a literal string
val rel2 = os.rel / "folder/file"

// The path "file"
val rel3 = os.rel / "file"

// The relative difference between two paths
val target = os.pwd / "target/file"
assert((target.relativeTo(os.pwd)) == os.rel / "target/file")

// `up`s get resolved automatically
val minus = os.pwd.relativeTo(target)
val ups = os.up / os.up
assert(minus == ups)

In general, very few APIs take relative paths. Their main purpose is to be combined with absolute paths in order to create new absolute paths. e.g.

val target = os.pwd / "target/file"
val difference = target.relativeTo(os.pwd)
val newBase = os.root / "code/server"
assert(newBase / difference == os.root / "code/server/target/file")

os.up is a relative path that comes in-built:

val target = os.root / "target/file"
assert(target / os.up == os.root / "target")

Note that all paths, both relative and absolute, are always expressed in a canonical manner:

assert((os.root / "folder/file" / os.up).toString == "/folder")
// not "/folder/file/.."

assert((os.rel / "folder/file" / os.up).toString == "folder")
// not "folder/file/.."

So you don’t need to worry about canonicalizing your paths before comparing them for equality or otherwise manipulating them.

`os.SubPath`

os.SubPaths represent relative paths without any .. segments. These are basically defined as:

class SubPath private[ops] (segments0: Array[String])

They can be created in the following ways:

// The path "folder/file" in multiple segments
val sub1 = os.sub / "folder" / "file"
// RHS of `/` can have multiple segments if-and-only-if it is a literal string
val sub2 = os.sub / "folder/file"

// The relative difference between two paths
val target = os.pwd / "out/scratch/file"
assert((target subRelativeTo os.pwd) == os.sub / "out/scratch/file")

// Converting os.RelPath to os.SubPath
val rel3 = os.rel / "folder/file"
val sub3 = rel3.asSubPath

os.SubPaths are useful for representing paths within a particular folder or directory. You can combine them with absolute os.Paths to resolve paths within them, without needing to worry about Directory Traversal Attacks du to accidentally accessing paths outside the destination folder.

val target = os.pwd / "target/file"
val difference = target.relativeTo(os.pwd)
val newBase = os.root / "code/server"
assert(newBase / difference == os.root / "code/server/target/file")

Attempting to construct an os.SubPath with .. segments results in an exception being thrown:

val target = os.pwd / "out/scratch" /

// `up`s are not allowed in sub paths
intercept[Exception](os.pwd subRelativeTo target)

Like os.Paths and os.RelPath, os.SubPaths are always canonicalized and can be compared for equality without worrying about different representations.

Path Operations

OS-Lib’s paths are transparent data-structures, and you can always access the segments and ups directly. Nevertheless, OS-Lib defines a number of useful operations that handle the common cases of dealing with these paths:

In this definition, ThisType represents the same type as the current path; e.g. a Path’s / returns a Path while a RelPath’s / returns a RelPath. Similarly, you can only compare or subtract paths of the same type.

Apart from os.RelPaths themselves, a number of other data structures are convertible into os.RelPaths when spliced into a path using /:

Strings
Symbols
Array[T]s where T is convertible into a RelPath
Seq[T]s where T is convertible into a RelPath

Constructing Paths

Apart from built-ins like os.pwd or os.root or os.home, you can also construct Paths from Strings, java.io.Files or java.nio.file.Paths:

val relStr = "hello/cow/world/.."
val absStr = "/hello/world"

assert(
  RelPath(relStr) == "hello/cow",
  // Path(...) also allows paths starting with ~,
  // which is expanded to become your home directory
  Path(absStr) == os.root / "hello/world"
)

// You can also pass in java.io.File and java.nio.file.Path
// objects instead of Strings when constructing paths
val relIoFile = new java.io.File(relStr)
val absNioFile = java.nio.file.Paths.get(absStr)

assert(
  RelPath(relIoFile) ==  "hello/cow",
  Path(absNioFile) == os.root / "hello/world",
  Path(relIoFile, root / "base") == os.root / "base/hello/cow"
)

Trying to construct invalid paths fails with exceptions:

val relStr = "hello/.."
intercept[java.lang.IllegalArgumentException]{
  Path(relStr)
}

val absStr = "/hello"
intercept[java.lang.IllegalArgumentException]{
  RelPath(absStr)
}

val tooManyUpsStr = "/hello/../.."
intercept[PathError.AbsolutePathOutsideRoot.type]{
  Path(tooManyUpsStr)
}

As you can see, attempting to parse a relative path with os.Path or an absolute path with os.RelPath throws an exception. If you’re uncertain about what kind of path you are getting, you could use BasePath to parse it :

val relStr = "hello/cow/world/.."
val absStr = "/hello/world"
assert(
  FilePath(relStr) == "hello/cow",
  FilePath(absStr) == os.root / "hello/world"
)

This converts it into a BasePath, which is either a os.Path or os.RelPath. It’s then up to you to pattern-match on the types and decide what you want to do in each case.

You can also pass in a second argument to Path(..., base). If the path being parsed is a relative path, this base will be used to coerce it into an absolute path:

val relStr = "hello/cow/world/.."
val absStr = "/hello/world"
val basePath: FilePath = FilePath(relStr)
assert(
  os.Path(relStr,   os.root / "base") == os.root / "base/hello/cow",
  os.Path(absStr,   os.root / "base") == os.root / "hello/world",
  os.Path(basePath, os.root / "base") == os.root / "base/hello/cow",
  os.Path(".", os.pwd).last != ""
)

For example, if you wanted the common behavior of converting relative paths to absolute based on your current working directory, you can pass in os.pwd as the second argument to Path(...). Apart from passing in Strings or java.io.Files or java.nio.file.Paths, you can also pass in BasePaths you parsed early as a convenient way of converting it to a absolute path, if it isn’t already one.

In general, OS-Lib is very picky about the distinction between relative and absolute paths, and doesn’t allow "automatic" conversion between them based on current-working-directory the same way many other filesystem APIs (Bash, Java, Python, …) do. Even in cases where it’s uncertain, e.g. you’re taking user input as a String, you have to either handle both possibilities with BasePath or explicitly choose to convert relative paths to absolute using some base.

Roots and filesystems

If you are using a system that supports different roots of paths, e.g. Windows, you can use the argument of os.root to specify which root you want to use. If not specified, the default root will be used (usually, C on Windows, / on Unix).

val root = os.root('C:\') / "Users/me"
assert(root == os.Path("C:\Users\me"))

Additionally, custom filesystems can be specified by passing a FileSystem to os.root. This allows you to use OS-Lib with non-standard filesystems, such as jar filesystems or in-memory filesystems.

val uri = new URI("jar", Paths.get("foo.jar").toURI().toString, null);
val env = new HashMap[String, String]();
env.put("create", "true");
val fs = FileSystems.newFileSystem(uri, env);
val path = os.root("/", fs) / "dir"

Note that the jar file system operations suchs as writing to a file are supported only on JVM 11+. Depending on the filesystem, some operations may not be supported - for example, running an os.proc with pwd in a jar file won’t work. You may also meet limitations imposed by the implementations - in jar file system, the files are created only after the file system is closed. Until that, the ones created in your program are kept in memory.

`os.ResourcePath`

In addition to manipulating paths on the filesystem, you can also manipulate os.ResourcePath in order to read resources off of the Java classpath. By default, the path used to load resources is absolute, using the Thread.currentThread().getContextClassLoader.

val contents = os.read(os.resource / "test/ammonite/ops/folder/file.txt")
assert(contents.contains("file contents lols"))

You can also pass in a classloader explicitly to the resource call:

val cl = getClass.getClassLoader
val contents2 = os.read(os.resource(cl)/ "test/ammonite/ops/folder/file.txt")
assert(contents2.contains("file contents lols"))

If you want to load resources relative to a particular class, pass in a class for the resource to be relative, or getClass to get something relative to the current class.

val cls = classOf[test.os.Testing]
val contents = os.read(os.resource(cls) / "folder/file.txt")
assert(contents.contains("file contents lols"))

val contents2 = os.read(os.resource(getClass) / "folder/file.txt")
assert(contents2.contains("file contents lols"))

In both cases, reading resources is performed as if you did not pass a leading slash into the getResource("foo/bar") call. In the case of ClassLoader#getResource, passing in a leading slash is never valid, and in the case of Class#getResource, passing in a leading slash is equivalent to calling getResource on the ClassLoader.

OS-Lib ensures you only use the two valid cases in the API, without a leading slash, and not the two cases with a leading slash which are redundant (in the case of Class#getResource, which can be replaced by ClassLoader#getResource) or invalid (a leading slash with ClassLoader#getResource)

Note that you can only use os.read from resource paths; you can’t write to them or perform any other filesystem operations on them, since they’re not really files.

Note also that resources belong to classloaders, and you may have multiple classloaders in your application e.g. if you are running in a servlet or REPL. Make sure you use the correct classloader (or a class belonging to the correct classloader) to load the resources you want, or else it might not find them.

`os.Source`

Many operations in OS-Lib operate on os.Sources. These represent values that can provide data which you can then use to write, transmit, etc.

By default, the following types of values can be used where-ever os.Sources are required:

Any geny.Writable data type:
- Array[Byte]
- java.lang.String (these are treated as UTF-8)
- java.io.InputStream
java.nio.channels.SeekableByteChannel
Any TraversableOnce[T] of the above: e.g. Seq[String], List[Array[Byte]], etc.

Some operations only work on os.SeekableSource, because they need the ability to seek to specific offsets in the data. Only the following types of values can be used where os.SeekableSource is required:

java.nio.channels.SeekableByteChannel

os.Source also supports anything that implements the Writable interface, such as ujson.Values, uPickle's upickle.default.writable values, or Scalatags's Tags

You can also convert an os.Path or os.ResourcePath to an os.Source via .toSource.

`os.Generator`

Taken from the geny library, os.Generators are similar to iterators except instead of providing:

def hasNext(): Boolean
def next(): T

os.Generators provide:

def generate(handleItem: A => Generator.Action): Generator.Action

In general, you should not notice much of a difference using Generators vs using Iterators: you can use the same .map/.filter/.reduce/etc. operations on them, and convert them to collections via the same .toList/.toArray/etc. conversions. The main difference is that Generators can enforce cleanup after traversal completes, so we can ensure open files are closed and resources are released without any accidental leaks.

`os.PermSet`

os.PermSets represent the filesystem permissions on a single file or folder. Anywhere an os.PermSet is required, you can pass in values of these types:

java.lang.Strings of the form "rw-r-xrwx", with r/w/x representing the permissions that are present or dashes - representing the permissions which are absent
Octal Ints of the form Integer.parseInt("777", 8), matching the octal 755 or 666 syntax used on the command line
Set[PosixFilePermission]

In places where os.PermSets are returned to you, you can then extract the string, int or set representations of the os.PermSet via:

perms.toInt(): Int
perms.toString(): String
perms.value: Set[PosixFilePermission]

Changelog

0.10.7

Allow multi-segment paths segments for literals com-lihaoyi#297: You can now write os.pwd / "foo/bar/qux" rather than os.pwd / "foo" / "bar" / "qux". Note that this is only allowed for string literals, and non-literal path segments still need to be wrapped e.g. def myString = "foo/bar/qux"; os.pwd / os.SubPath(myString) for security and safety purposes

0.10.6

Make os.pwd modifiable via the os.dynamicPwd dynamic variable com-lihaoyi#298

0.10.5

Introduce os.SubProcess.env DynamicVariable to override default env (com-lihaoyi#295)

0.10.4

Add a lightweight syntax for os.call() and os.spawn APIs (com-lihaoyi#292)
Add a configurable grace period when subprocesses timeout and have to be terminated to give a chance for shutdown logic to run (com-lihaoyi#286)

0.10.3

os.Inherit now can be redirected on a threadlocal basis via os.Inherit.in, .out, or .err. os.InheritRaw is available if you do not want the redirects to take effect

0.10.2

Support os.proc on Scala Native (com-lihaoyi#257)

0.10.1

Fix os.copy and os.move directories to root (#267)

0.10.0

Support for Scala-Native 0.5.0
Dropped support for Scala 2.11.x
Minimum version of Scala 3 increased to 3.3.1

0.9.3 - 2024-01-01

Fix os.watch on Windows (#236)
Fix propagateEnv = false to not propagate env (#238)
Make os.home a def (#239)

0.9.2 - 2023-11-05

Added new convenience API to create pipes between processes with .pipeTo
Fixed issue with leading .. / os.up in path segments created from a Seq
Fixed Windows-specific issues with relative paths with leading (back)slashes
Removed some internal use of deprecated API
ScalaDoc now maps some external references to their online sites
Dependency updates: sourcecode 0.3.1
Tooling updates: acyclic 0.3.9, Mill 0.11.5, mill-mima 0.0.24, mill-vcs-version 0.4.0, scalafmt 3.7.15

0.9.1 - 2023-03-07

Refined return types when constructing paths with / and get rid of long ThisType#ThisType cascades.
Added a new PathConvertible to support `URI`s when constructing paths.

0.9.0 - 2022-11-28

os.proc now also supports CharSequence(s) as Shellable
ProcessResult now also contains the actual used command
Fixed handling of atime and ctime in StatInfo
Deleted ConcurrentLinkedQueue from Scala Native jars, as it is now provided by Scala Native 0.4 itself
Enabled MiMa checks to CI setup and officially support early semantic versioning since this release
Documentation improvements

Older releases

0.8.1 - 2022-01-31

Added support for Scala Native on Scala 3

0.8.0 - 2021-12-11

Avoid throwing an exception when sorting identical paths #90
Make os.remove behave more like Files.deleteIfExists #89
Make .ext on empty paths return "" rather than crashing #87

0.7.8 - 2021-05-27

Restored binary compatibility in os.copy and os.copy.into to os-lib versions before 0.7.5

0.7.7 - 2021-05-14

Add support for Scala 3.0.0

0.7.6 - 2021-04-28

Add support for Scala 3.0.0-RC3

0.7.5 - 2021-04-21

Re-added support for Scala 2.11
Added new option mergeFolders to os.copy
os.copy now honors followLinks when copying symbolic links to directories

0.7.4

Add support for Scala 3.0.0-RC2

0.7.3

Add support for Scala 3.0.0-RC1
Migration of the CI system from Travis CI to GitHub Actions

0.7.2

Add support for Scala 3.0.0-M3

0.7.1

Improve performance of os.write by buffering output stream to files

0.6.2

Moved the os.Bytes, os.StreamValue (now named ByteData) interfaces into geny package, for sharing with Requests-Scala
Add os.read.stream function, that returns a geny.Readable

0.5.0

os.Source now supports any data type that is geny.Writable

0.4.2

Added a new os.SubPath data type, for safer handling of sub-paths within a directory.
Removed os.proc.stream, since you can now customize the stdout or stderr of os.proc.call to handle output in a streaming fashion
stderr in os.proc.call and os.proc.spawn defaults to os.Inherit rather than os.Pipe; pass in stderr = os.Pipe explicitly to get back the old behavior
Fix timeout not working with os.proc.call #27
Attempt to fix crasher accessing os.pwd #24
Added an os-lib-watch package, which can be used to efficiently recursively watch folders for updates #23
os.stat no longer provides POSIX owner/permissions related metadata by default #15, use os.stat.posix to fetch that separately
os.stat.full has been superseded by os.stat and os.stat.posix
Removed os.BasicStatInfo, which has been superseded by os.StatInfo

0.3.0

Support for Scala 2.13.0 final

0.2.8

os.ProcessOutput trait is no longer sealed

0.2.7

Narrow return type of readLink.absolute from FilePath to Path
Fix handling of standaline \r in os.SubProcess#stdout.readLine

0.2.6

Remove os.StatInfo#name, os.BasicStatInfo#name and os.FullStatInfo#name, since it is just the last path segment of the stat call and doesn’t properly reflect the actual name of the file on disk (e.g. on case-insensitive filesystems)
os.walk.attrs and os.walk.stream.attrs now provides a os.BasicFileInfo to the skip predicate.
Add os.BasePath#baseName, which returns the section of the path before the os.BasePath#ext extension.

0.2.5

New os.readLink/os.readLink.absolute methods to read the contents of symbolic links without dereferencing them.
New os.read.chunked(p: Path, chunkSize: Int): os.Generator[(Array[Byte], Int)] method for conveniently iterating over chunks of a file
New os.truncate(p: Path, size: Int) method
SubProcess streams now implement java.io.DataInput/DataOutput for convenience
SubProcess streams are now synchronized for thread-safety
os.write now has createFolders default to false
os.Generator now has a .withFilter method
os.symlink now allows relative paths
os.remove.all now properly removes broken symlinks, and no longer recurses into the symlink’s contents
os.SubProcess now implements java.lang.AutoCloseable
New write.channel counterpart to read.channel (and write.over.channel and write.append.channel)
os.PermSet is now modelled internally as a boxed Int for performance, and is a case class with proper equals/hashcode
os.read.bytes(arg: Path, offset: Long, count: Int) no longer leaks open file channels
Reversed the order of arguments in os.symlink and os.hardlink, to match the order of the underlying java NIO functions.

0.2.2

Allow chaining of multiple subprocesses stdin/stdout

0.2.0

First release

Name		Name	Last commit message	Last commit date
Latest commit History 413 Commits
.github		.github
os		os
.editorconfig		.editorconfig
.git-blame-ignore-revs		.git-blame-ignore-revs
.gitignore		.gitignore
.mill-version		.mill-version
.scalafmt.conf		.scalafmt.conf
LICENSE		LICENSE
Readme.adoc		Readme.adoc
build.sc		build.sc
mill		mill

License

wb14123/os-lib

Folders and files

Latest commit

History

Repository files navigation

OS-Lib

Getting Started

Cookbook

Concatenate text files

Spawning a subprocess on multiple files

Curl URL to temporary file

Recursive line count

Largest Three Files

Moving files out of folder

Calculate word frequencies

Operations

Reading & Writing

os.read

os.read.bytes

os.read.chunks

os.read.lines

os.read.lines.stream

os.read.inputStream

os.read.stream

os.write

os.write.append

os.write.over

os.write.outputStream

os.truncate

Listing & Walking

os.list

os.list.stream

os.walk

os.walk.attrs

os.walk.stream

os.walk.stream.attrs

Manipulating Files & Folders

os.exists

os.move

os.move.matching

os.move.into

os.move.over

os.copy

os.copy.into

os.copy.over

os.copy with mergeFolders

os.makeDir

os.makeDir.all

os.remove

os.remove.all

os.hardlink

os.symlink

os.readLink

os.followLink

os.temp

os.temp.dir