better-files License TravisCI Codacy

better-files is a dependency-free pragmatic thin Scala wrapper around Java NIO.

Talks Gitter

ScalaDays NYC 2016: Introduction to better-files

Tutorial Scaladoc

  1. Instantiation
  2. Simple I/O
  3. Resource APIs
  4. Streams
  5. Encodings
  6. Java serialization utils
  7. Java compatibility
  8. Globbing
  9. File system operations
  10. Temporary files
  11. UNIX DSL
  12. File attributes
  13. File comparison
  14. Zip/GZip
  15. Automatic Resource Management
  16. Scanner
  17. File Monitoring
  18. Reactive File Watcher

sbt UpdateImpact

In your build.sbt, add this:

libraryDependencies += "com.github.pathikrit" %% "better-files" % version

To use the Akka based file monitor, also add this:

libraryDependencies ++= Seq(
  "com.github.pathikrit"  %% "better-files-akka"  % version,
  "com.typesafe.akka"     %% "akka-actor"         % "2.5.13"
)

Latest version: Maven Scaladex

Although this library is currently only actively developed for Scala 2.11, 2.12 and 2.13, you can find reasonably recent versions of this library for Scala 2.10 here.

Tests codecov


Instantiation

The following are all equivalent:

import better.files._
import File._
import java.io.{File => JFile}

val f = File("/User/johndoe/Documents")                      // using constructor
val f1: File = file"/User/johndoe/Documents"                 // using string interpolator
val f2: File = "/User/johndoe/Documents".toFile              // convert a string path to a file
val f3: File = new JFile("/User/johndoe/Documents").toScala  // convert a Java file to Scala
val f4: File = root/"User"/"johndoe"/"Documents"             // using root helper to start from root
val f5: File = `~` / "Documents"                             // also equivalent to `home / "Documents"`
val f6: File = "/User"/"johndoe"/"Documents"                 // using file separator DSL
val f7: File = "/User"/'johndoe/'Documents                   // same as above but using Symbols instead of Strings
val f8: File = home/"Documents"/"presentations"/`..`         // use `..` to navigate up to parent

Note: Rename the import if you think the usage of the class File may confuse your teammates:

import better.files.{File => ScalaFile, _}
import java.io.File

I personally prefer renaming the Java crap instead:

import better.files._
import java.io.{File => JFile}

File Read/Write

Dead simple I/O:

val file = root/"tmp"/"test.txt"
file.overwrite("hello")
file.appendLine().append("world")
assert(file.contentAsString == "hello\nworld")

If you are someone who likes symbols, then the above code can also be written as:

import better.files.Dsl.SymbolicOperations

file < "hello"     // same as file.overwrite("hello")
file << "world"    // same as file.appendLines("world")
assert(file! == "hello\nworld")

Or even, right-associatively:

import better.files.Dsl.SymbolicOperations

"hello" `>:` file
"world" >>: file
val bytes: Array[Byte] = file.loadBytes

Fluent Interface:

 (root/"tmp"/"diary.txt")
  .createIfNotExists()
  .appendLine()
  .appendLines("My name is", "Inigo Montoya")
  .moveToDirectory(home/"Documents")
  .renameTo("princess_diary.txt")
  .changeExtensionTo(".md")
  .lines

Resource APIs

Confused by the various ways to load resources in Java? Worry no more:

val resource        : InputStream   = Resource.getAsStream("foo.txt") //Same as this.getClass().getResource("foo.txt")
val resourceURL     : java.net.URL  = Resource.getUrl("foo.txt")
val resourceAsStr   : String        = Resource.getAsString("foo.txt")

Tha above APIs can load from custom ClassLoaders too:

val resource        : InputStream   = Resource.at[MyClass].getAsStream("foo.txt")

Streams

Various ways to slurp a file without loading the contents into memory:

val bytes  : Iterator[Byte]            = file.bytes
val chars  : Iterator[Char]            = file.chars
val lines  : Iterator[String]          = file.lineIterator      //file.lines loads all lines in memory

Note: The above APIs can be traversed at most once e.g. file.bytes is a Iterator[Byte] which only allows TraversableOnce. To traverse it multiple times without creating a new iterator instance, convert it into some other collection e.g. file.bytes.toStream

You can write an Iterator[Byte] or an Iterator[String] back to a file:

file.writeBytes(bytes)
file.printLines(lines)

tee multiple outputstreams:

val s3 = s1 tee s2
s3.printWriter.println(s"Hello world") // gets written to both s1 and s2

Encodings

You can supply your own charset too for anything that does a read/write (it assumes java.nio.charset.Charset.defaultCharset() if you don't provide one):

val content: String = file.contentAsString  // default charset

// custom charset:
import java.nio.charset.Charset
file.contentAsString(charset = Charset.forName("US-ASCII"))

//or simply using implicit conversion from Strings
file.write("hello world")(charset = "US-ASCII")

Note: By default, better-files correctly handles BOMs while decoding. If you wish to have the incorrect JDK behaviour, you would need to supply Java's UTF-8 charset e.g.:

file.contentAsString(charset = Charset.forName("UTF-8"))    // Default incorrect JDK behaviour for UTF-8 (see: JDK-4508058)

If you also wish to write BOMs while encoding, you would need to supply it as:

file.write("hello world")(charset = UnicodeCharset("UTF-8", writeByteOrderMarkers = true))

Java serialization utils

Some common utils to serialize/deserialize using Java's serialization util

case class Person(name: String, age: Int)
val person = new Person("Chris", 24)

// Write
file.newOutputStream.asObjectOutputStream.serialize(obj).flush()

// Read
val person2 = file.newInputStream.asObjectInputStream.deserialize[Person]
assert(person == person2)

// Read using custom class loader:
file.newInputStream.asObjectInputStreamUsingClassLoader(classLoader = myClassLoader).deserialize[Person]

The above can be simply written as:

val person2: Person = file.writeSerialized(person).readDeserialized[Person]
assert(person == person2)

Java interoperability

You can always access the Java I/O classes:

val file: File = tmp / "hello.txt"
val javaFile     : java.io.File                 = file.toJava
val uri          : java.net.URI                 = file.uri
val url          : java.net.URL                 = file.url
val reader       : java.io.BufferedReader       = file.newBufferedReader
val outputstream : java.io.OutputStream         = file.newOutputStream
val writer       : java.io.BufferedWriter       = file.newBufferedWriter
val inputstream  : java.io.InputStream          = file.newInputStream
val path         : java.nio.file.Path           = file.path
val fs           : java.nio.file.FileSystem     = file.fileSystem
val channel      : java.nio.channel.FileChannel = file.newFileChannel
val ram          : java.io.RandomAccessFile     = file.newRandomAccess
val fr           : java.io.FileReader           = file.newFileReader
val fw           : java.io.FileWriter           = file.newFileWriter(append = true)
val printer      : java.io.PrintWriter          = file.newPrintWriter

The library also adds some useful implicits to above classes e.g.:

file1.reader > file2.writer       // pipes a reader to a writer
System.in > file2.out             // pipes an inputstream to an outputstream
src.pipeTo(sink)                  // if you don't like symbols

val bytes   : Iterator[Byte]        = inputstream.bytes
val bis     : BufferedInputStream   = inputstream.buffered
val bos     : BufferedOutputStream  = outputstream.buffered
val reader  : InputStreamReader     = inputstream.reader
val writer  : OutputStreamWriter    = outputstream.writer
val printer : PrintWriter           = outputstream.printWriter
val br      : BufferedReader        = reader.buffered
val bw      : BufferedWriter        = writer.buffered
val mm      : MappedByteBuffer      = fileChannel.toMappedByteBuffer
val str     : String                = inputstream.asString  //Read a string from an InputStream
val in      : InputStream           = str.inputStream
val reader  : Reader                = str.reader
val lines   : Seq[String]           = str.lines

better-files also supports certain conversions that are not supported out of the box by the JDK

Globbing

No need to port this to Scala:

val dir = "src"/"test"
val matches: Iterator[File] = dir.glob("*.{java,scala}")
// above code is equivalent to:
dir.listRecursively.filter(f => f.extension == Some(".java") || f.extension == Some(".scala"))

You can even use more advanced regex syntax instead of glob syntax:

val matches = dir.globRegex("^\\w*$".r) //equivalent to dir.glob("^\\w*$")(syntax = File.PathMatcherSyntax.regex)

By default, glob syntax in better-files is different from the default JDK glob behaviour since it always includes path. To use the default behaviour:

dir.glob("**/*.txt", includePath = false) // JDK default
//OR
dir.glob("*.txt", includePath = true) // better-files default

You can also extend the File.PathMatcherSyntax to create your own matchers.

For custom cases:

dir.collectChildren(_.isSymbolicLink) // collect all symlinks in a directory

For simpler cases, you can always use dir.list or dir.walk(maxDepth: Int)

File system operations

Utilities to ls, cp, rm, mv, ln, md5, touch, cat etc:

file.touch()
file.delete()     // unlike the Java API, also works on directories as expected (deletes children recursively)
file.clear()      // If directory, deletes all children; if file clears contents
file.renameTo(newName: String)
file.moveTo(destination)
file.moveToDirectory(destination)
file.copyTo(destination)       // unlike the default API, also works on directories (copies recursively)
file.copyToDirectory(destination)
file.linkTo(destination)                     // ln destination file
file.symbolicLinkTo(destination)             // ln -s destination file
file.{checksum, md5, sha1, sha256, sha512, digest}   // also works for directories
file.setOwner(user: String)      // chown user file
file.setGroup(group: String)     // chgrp group file
Seq(file1, file2) `>:` file3     // same as cat file1 file2 > file3 (must import import better.files.Dsl.SymbolicOperations)
Seq(file1, file2) >>: file3      // same as cat file1 file2 >> file3 (must import import better.files.Dsl.SymbolicOperations)
file.isReadLocked; file.isWriteLocked; file.isLocked
File.numberOfOpenFileDescriptors        // number of open file descriptors

Temporary files

Utils to create temporary files:

File.newTemporaryDirectory()
File.newTemporaryFile()

The above APIs allow optional specifications of prefix, suffix and parentDir. These files are not deleted automatically on exit by the JVM (you have to set deleteOnExit which adds to shutdownHook).

A cleaner alternative is to use self-deleting file contexts which deletes the file immediately when done:

for {
 tempFile <- File.temporaryFile()
} doSomething(tempFile) // tempFile is auto deleted at the end of this block - even if an exception happens

OR equivalently:

File.usingTemporaryFile() {tempFile =>
  //do something
}  // tempFile is auto deleted at the end of this block - even if an exception happens

You can make any files temporary (i.e. delete after use) by doing this:

val foo = File.home / "Downloads" / "foo.txt"

for {
 temp <- foo.toTemporary
} doSomething(temp) // foo is deleted at the end of this block - even if an exception happens

UNIX DSL

All the above can also be expressed using methods reminiscent of the command line:

import better.files._
import better.files.Dsl._   // must import Dsl._ to bring in these utils

pwd / cwd     // current dir
cp(file1, file2)
mv(file1, file2)
rm(file) /*or*/ del(file)
ls(file) /*or*/ dir(file)
ln(file1, file2)     // hard link
ln_s(file1, file2)   // soft link
cat(file1)
cat(file1) >>: file
touch(file)
mkdir(file)
mkdirs(file)         // mkdir -p
chown(owner, file)
chgrp(owner, file)
chmod_+(permission, files)  // add permission
chmod_-(permission, files)  // remove permission
md5(file); sha1(file); sha256(file); sha512(file)
unzip(zipFile)(targetDir)
zip(file*)(targetZipFile)
ungzip(gzipFile)(targetFile)
gzip(file)(targetGZipFile)

File attributes

Query various file attributes e.g.:

file.name       // simpler than java.io.File#getName
file.extension
file.contentType
file.lastModifiedTime     // returns JSR-310 time
file.owner
file.group
file.isDirectory; file.isSymbolicLink; file.isRegularFile
file.isHidden
file.hide(); file.unhide()
file.isOwnerExecutable; file.isGroupReadable // etc. see file.permissions
file.size                 // for a directory, computes the directory size
file.posixAttributes; file.dosAttributes  // see file.attributes
file.isEmpty      // true if file has no content (or no children if directory) or does not exist
file.isParentOf; file.isChildOf; file.isSiblingOf; file.siblings
file("dos:system") = true  // set custom meta-data for file (similar to Files.setAttribute)

All the above APIs let you specify the LinkOption either directly:

file.isDirectory(LinkOption.NOFOLLOW_LINKS)

Or using the File.LinkOptions helper:

file.isDirectory(File.LinkOptions.noFollow)

chmod:

import java.nio.file.attribute.PosixFilePermission
file.addPermission(PosixFilePermission.OWNER_EXECUTE)      // chmod +X file
file.removePermission(PosixFilePermission.OWNER_WRITE)     // chmod -w file
assert(file.permissionsAsString == "rw-r--r--")

// The following are all equivalent:
assert(file.permissions contains PosixFilePermission.OWNER_EXECUTE)
assert(file.testPermission(PosixFilePermission.OWNER_EXECUTE))
assert(file.isOwnerExecutable)

File comparison

Use == to check for path-based equality and === for content-based equality:

file1 == file2    // equivalent to `file1.isSamePathAs(file2)`
file1 === file2   // equivalent to `file1.isSameContentAs(file2)` (works for regular-files and directories)
file1 != file2    // equivalent to `!file1.isSamePathAs(file2)`
file1 !== file2   // equivalent to `!file1.isSameContentAs(file2)`

There are also various Ordering[File] instances included, e.g.:

val files = myDir.list.toSeq
files.sorted(File.Order.byName)
files.max(File.Order.bySize)
files.min(File.Order.byDepth)
files.max(File.Order.byModificationTime)
files.sorted(File.Order.byDirectoriesFirst)

Zip APIs

You don't have to lookup on StackOverflow "How to zip/unzip/gzip in Java/Scala?":

// Unzipping:
val zipFile: File = file"path/to/research.zip"
val research: File = zipFile.unzipTo(destination = home/"Documents"/"research")

// Zipping:
val zipFile: File = directory.zipTo(destination = home/"Desktop"/"toEmail.zip")

// Zipping in:
val zipFile = File("countries.zip").zipIn(Iterator(file"usa.txt", file"russia.txt"))()

// Zipping/Unzipping to temporary files/directories:
val someTempZipFile: File = directory.zip()
val someTempDir: File = someTempZipFile.unzip()
assert(directory === someTempDir)

GZIP handling:

File("big-data.csv").gzipTo(File("big-data.csv.gz"))
File("big-data.csv.gz").unGzipTo(File("big-data.csv"))

// GZIP stream handling:
File("countries.gz").newInputStream.asGzipInputStream().lines.take(10).foreach(println)

def write(out: OutputStream, countries: Seq[String]) =
  out.asGzipOutputStream().printWriter().printLines(countries).close()

Lightweight ARM

Auto-close Java closeables:

for {
  in <- file1.newInputStream.autoClosed
  out <- file2.newOutputStream.autoClosed
} in.pipeTo(out)
// The input and output streams are auto-closed once out of scope

better-files provides convenient managed versions of all the Java closeables e.g. instead of writing:

for {
 reader <- file.newBufferedReader.autoClosed
} foo(reader)

You can write:

for {
 reader <- file.bufferedReader    // returns Dispose[BufferedReader]
} foo(reader)

// or simply:
file.bufferedReader.foreach(foo)

Similarly:

for {
 reader <- file.bufferedReader
} yield foo(reader)

// Simpler
file.bufferedReader.map(foo).get()

// Even simpler
file.bufferedReader.apply(foo)

If foo itself is lazy and depends on reader being open, you should flatMap instead of apply:

def lines(reader: BufferedReader): Iterator[String] = ???

for {
  reader <- file.bufferedReader
  line <- lines(reader)
} yield line

// or simply
file.bufferedReader.flatMap(lines)

You can also define your own custom disposable resources e.g.:

trait Shutdownable {
  def shutdown(): Unit = ()
}

object Shutdownable {
  implicit val disposable: Disposable[Shutdownable] = Disposable(_.shutdown())
}

val s: Shutdownable = ....

for {
  instance <- new Dispose(s)
} doSomething(s)  // s is disposed after this

Scanner

Although java.util.Scanner has a feature-rich API, it only allows parsing primitives. It is also notoriously slow since it uses regexes and does un-Scala things like returns nulls and throws exceptions.

better-files provides a faster, richer, safer, more idiomatic and compossible Scala replacement that does not use regexes, allows peeking, accessing line numbers, returns Options whenever possible and lets the user mixin custom parsers:

val f1 = File("/tmp/temp.txt")
val data = f1.overwrite(s"""Hello World
  | 1 true
  | 2 3
""".stripMargin)
val scanner: Scanner = data.newScanner()
assert(scanner.next[String] == "Hello")
assert(scanner.lineNumber == 1)
assert(scanner.next[String] == "World")
assert(scanner.next[(Int, Boolean)] == (1, true))
assert(scanner.nextLine() == " 2 3")
assert(!scanner.hasNext)

If you are simply interested in tokens, you can use file.tokens()

Writing your own custom scanners:

sealed trait Animal
case class Dog(name: String) extends Animal
case class Cat(name: String) extends Animal

implicit val animalParser: Scannable[Animal] = Scannable {scanner =>
  val name = scanner.next[String]
  if (name == "Garfield") Cat(name) else Dog(name)
}

val scanner = file.newScanner()
println(scanner.next[Animal])

The shapeless-scanner lets you scan HLists:

val in = Scanner("""
  12 Bob True
  13 Mary False
  26 Rick True
""")

import shapeless._

type Row = Int :: String :: Boolean :: HNil

val out = Seq.fill(3)(in.next[Row])
assert(out == Seq(
  12 :: "Bob" :: true :: HNil,
  13 :: "Mary" :: false :: HNil,
  26 :: "Rick" :: true :: HNil
))

and case-classes:

case class Person(id: Int, name: String, isMale: Boolean)
val out2 = Seq.fill(3)(in.next[Person])

Simple CSV reader:

val file = """
  23,foo
  42,bar
"""
val csvScanner = file.newScanner(StringSpliiter.on(','))
csvScanner.next[Int]    //23
csvScanner.next[String] //foo

File Monitoring

Vanilla Java watchers:

import java.nio.file.{StandardWatchEventKinds => EventType}
val service: java.nio.file.WatchService = myDir.newWatchService
myDir.register(service, events = Seq(EventType.ENTRY_CREATE, EventType.ENTRY_DELETE))

The above APIs are cumbersome to use (involves a lot of type-casting and null-checking), are based on a blocking polling-based model, does not easily allow recursive watching of directories and nor does it easily allow watching regular files without writing a lot of Java boilerplate.

better-files abstracts all the above ugliness behind a simple interface:

val watcher = new FileMonitor(myDir, recursive = true) {
  override def onCreate(file: File, count: Int) = println(s"$file got created")
  override def onModify(file: File, count: Int) = println(s"$file got modified $count times")
  override def onDelete(file: File, count: Int) = println(s"$file got deleted")
}
watcher.start()

Sometimes, instead of overwriting each of the 3 methods above, it is more convenient to override the dispatcher itself:

import java.nio.file.{Path, StandardWatchEventKinds => EventType, WatchEvent}

val watcher = new FileMonitor(myDir, recursive = true) {
  override def onEvent(eventType: WatchEvent.Kind[Path], file: File, count: Int) = eventType match {
    case EventType.ENTRY_CREATE => println(s"$file got created")
    case EventType.ENTRY_MODIFY => println(s"$file got modified $count")
    case EventType.ENTRY_DELETE => println(s"$file got deleted")
  }
}

There is also an external module which gives high performance file monitoring and interpolates with better-files. See: https://github.com/gmethvin/directory-watcher#better-files-integration-scala

Akka File Watcher

better-files also provides a powerful yet concise reactive file watcher based on Akka actors that supports dynamic dispatches:

import akka.actor.{ActorRef, ActorSystem}
import better.files._, FileWatcher._

implicit val system = ActorSystem("mySystem")

val watcher: ActorRef = (home/"Downloads").newWatcher(recursive = true)

// register partial function for an event
watcher ! on(EventType.ENTRY_DELETE) {
 case file if file.isDirectory => println(s"$file got deleted")
}

// watch for multiple events
watcher ! when(events = EventType.ENTRY_CREATE, EventType.ENTRY_MODIFY) {
 case (EventType.ENTRY_CREATE, file, count) => println(s"$file got created")
 case (EventType.ENTRY_MODIFY, file, count) => println(s"$file got modified $count times")
}

Benchmarks

> sbt "core/testOnly better.files.benchmarks.*"
JavaScanner              : 2191 ms
StringBuilderScanner     : 1325 ms
CharBufferScanner        : 1117 ms
StreamingScanner         :  212 ms
IterableScanner          :  365 ms
IteratorScanner          :  297 ms
BetterFilesScanner       :  272 ms
ArrayBufferScanner       :  220 ms
FastJavaIOScanner2       :  181 ms
FastJavaIOScanner        :  179 ms

YourKit

YourKit supports better-files with its full-featured Java Profiler. YourKit, LLC is the creator of YourKit Java Profiler and YourKit .NET Profiler, innovative and intelligent tools for profiling Java and .NET applications.