layout | title |
---|---|
default |
ABSA Commons |
Selection of useful reusable components
import CollectionImplicits._
val iter: Iterator[_] = ???
val arr: Array[_] = ???
// copy 42 items from the `iter` to the `arr` with array offset 7
iter.fetchToArray(arr, 7, 42) // returns a number of actually copied items
A simple stackable Converter
trait with a simple memoized wrapper.
// 1. Define your converter
class AlchemicalConverter extends Converter {
override type From = Lead
override type To = Gold
override def convert(arg: Lead): Gold = ??? // treat with dragon eyes and cosmic rays
}
// 2. Instantiate it
val forge = new AlchemicalConverter
// or a memoized variant
val forge = new AlchemicalConverter with CachingConverter
// 3. Use it
forge.convert(... some lead ...) // returns some gold
Utility object that defines extended type constraints to be used in Scala type definitions.
In particular it defines a not
type constraint.
trait VegetarianMenu {
def add[A <: Food : not[Meat]#λ](food: A)
}
// Strings
(null: String).nonBlankOption // == None
" \t \n \r ".nonBlankOption // == None
"foo bar".nonBlankOption // == Some("foo bar")
// Collections
Seq.empty.asOption // == None
Seq(1, 2).asOption // == Some(Seq(1, 2))
// Just another way of doing Option(foo)
foo.asOption
// Optionally call a method in a chain
new MyBuilder
.withX(42)
.withY(77)
.optionally(_.withZ, maybeZ) // <---- withZ is called with a `value` if `maybeZ` is `Some(value)`
.optionally(_.withABC(a, _, c), maybeB) // <---- it also works with n-ary methods
conf.getRequiredXXX(...)
ReflectionUtils.directSubClassesOf[Food] // == Seq(classOf[Vegetables], classOf[Meat], classOf[Fish])
ReflectionUtils.objectsOf[Currency] // == Seq(classOf[EUR], classOf[USD], classOf[CZK])
ReflectionUtils.objectForName[MySingleton]("com.example.MySingleton") // == MySingleton
ReflectionUtils.extractFieldValue[Int](foo, "bar")
// or if you know a type where the field is declared
ReflectionUtils.extractFieldValue[Doh, Int](foo, "bar")
case class Person(name: String, age: Int, sex: Sex)
val aPerson = Person("Alex", 41, Male)
ReflectionUtils.extractProperties(aPerson) // == Map("name" -> "Alex, "age" -> 42, "sex" -> Male)
case class Button(title: String, isPressed = false)
ReflectionUtils.caseClassCtorArgDefaultValue[Int](classOf[Button], "name") // == None
ReflectionUtils.caseClassCtorArgDefaultValue[Int](classOf[Button], "isPressed") // == Some(false)
ReflectionUtils.ReflectionUtils.allInterfacesOf[A]
// or
ReflectionUtils.ReflectionUtils.allInterfacesOf(aClass)
Can be used to e.g. in a Case Object Enumeration pattern.
A similar solution and the motivation is well describes in the
Scala Enumerations hell article.
But unlike the above approach EnumerationMacros.sealedInstancesOf
utilizes Scala compiler macros,
so that the instances are lookup at the compile time.
sealed trait Color
object Color {
// returns Set(Red, Green, Blue)
val values: Set[Color] = EnumerationMacros.sealedInstancesOf[Color]
case object Red extends Color
case object Green extends Color
case object Blue extends Color
}
If you wants some code to be linked and executed at run-time, here's a simple way to do it:
val fn = ReflectionUtils.compile(q""" some scala code """)
fn()
Of with input parameters:
val fn = ReflectionUtils.compile(q"""
val foo = arg("foo")
val bar = arg("bar")
foo + bar
""")
fn(Map("foo" -> ..., "bar" -> ...))
Sometimes you need to support different versions of some library with a breaking changes in API. If there aren't too many breaking changes, or you only use a certain subset of an API then creating a proper adapter layer could be an overkill. In that case simple run-time evaluation could be a decent alternative:
object FilenameExtractor extends AccessorMethodValueExtractor[String]("filename", "name", "file")
val FilenameExtractor(fileName) = someObjectRepresentingAFile
// The first matching accessor name with type wins
If you try to pattern-match on a type that is missing from the classpath at runtime (e.g. optional dependency) you'll get NoClassDefFoundError
. Though it looks strange as from the use case perspective if there is no Foo
class there couldn't be a Foo
instance. Logically one would expect it to just not match, but in fact it throws.
aObject match {
...
case foo: Foo => // <----- this could throw NoClassDefFoundError !!
...
}
To get a desired behavior you can use SafeTypeMatchingExtractor
:
object FooExtractor extends SafeTypeMatchingExtractor("com.example.Foo")
aObject {
...
case FooExtractor(foo) => // do something with `foo`
...
}
or you can even make a fancy DSL from it:
object `_: Foo` extends SafeTypeMatchingExtractor("com.example.Foo")
aObject {
...
case `_: Foo`(foo) => // do something with `foo`
...
}
A singleton that parses build.properties
from the classpath and return version and timestamp as constants.
Copy build.properties.template
file and paste in into your classpath root as build.properties
. Make sure the resource filtering is enabled on your project build.
Then you can access it's content as simply as this:
BuildInfo.Version // returns `build.version` property from the `build.properties` file
BuildInfo.Timestamp // returns `build.timestamp` property from the `build.properties` file
BuildInfo.BuildProps // returns entire `build.properties` content as immutable Java `Properties`
If needed, you can customize a build.properties
resource path and/or the property mapping.
// loads '/foo/bar.properties' from the classpath
object MyBuildInfo extends BuildInfo(resourcePrefix = "/foo/bar")
object MyBuildInfo extends BuildInfo(propMapping = PropMapping(
version = "bld.ver", // binds "Version" field to "bld.ver" property
timestamp = "bld.ttt" // binds "Timestamp" field to "bld.ttt" property
))
You can also use apply()
method instead of inheritance. It all depends on your preferred code style:
val myBuildInfo = BuildInfo(...)
An easy way to create a temporary file or directory with the support for automatic recursive deletion (as rm -rf
) on JVM shutdown.
val myTmpFile = TempFile.deleteOnExit.path
val myTmpDir = TempDirectory.deleteOnExit.path
It also mimics Java IO API for a similar purpose
TempFile("myPrefix", "mySuffix")
A set of stackable traits, serving a wrapper around the way how Json4s (de)serializers are created. Instead of relying on implicit Formats
objects a stackable traits are used.
This API is also binary compatible to Json4s 3.2 and 3.3+ versions (Jackson and Native)
If you use Jackson impl and default formats, then you can simply do this:
class MyApp extends App with DefaultJacksonJsonSerDe {
FooBar.toJson // returns JSON string
"{...}".fromJson[FooBar] // returns a FooBar instance
}
Or you can create a singleton and use that instead:
object JsonSerDe extends DefaultJacksonJsonSerDe
import JsonSerDe._
fooBar.toJson
If you want another parser impl (e.g. Native), then you do this:
class MyApp extends App
with AbstractJsonSerDe
with native.JsonMethods
with DefaultFormatsBuilder {
...
fooBar.toJson
...
}
If you want custom formats than instead of mixing in DefaultFormatsBuilder
simply override def formats
method.
A simple utility that parses version strings. It supports SemVer 2.0 as well as a simple dot-separated version format. Can be used to compare the versions, for instance when implementing version predicates.
import Version._
require(Version.asSimple(SPARK_VERSION) > ver"2.4")
// or
require(Version.asSemVer(SomeLibVersion) > semver"1.2.3-beta.2")
To get a string representation from a Version
instance asString
extension method can be used:
val myVer: Version = semver"1.2.3-beta.2+777.42"
myVer.asString // returns "1.2.3-beta.2+777.42"
Semantic Versioning specific operations:
import Version._
val myVer = semver"111.222.333-alpha.444+build.555"
myVer.major == 111
myVer.minor == 222
myVer.patch == 333
myVer.core == semver"111.222.333"
myVer.preRelease == ver"alpha.444"
myVer.buildMeta == ver"build.555"
-
ConditionalTestTags
- runs certain tests conditionallyit should "test that new Spark feature" taggedAs ignoreIf(ver"$SPARK_VERSION" < ver"2.4") in { ... } it should "test some DAO" taggedAs ignoreIf(!isDatabaseAvailable) in { ... }
-
ConsoleStubs
- stubs console IOcaptureStdOut(Console.out.print("foo")) should be("foo")
-
SystemExitFixture
- interceptsSystem.exit()
and asserts statuscaptureExitStatus(System.exit(42)) should be(42) // OR assertingExitStatus(be > 0 and be < 5) { // run some code that calls System.exit(...) }
-
WhitespaceNormalizations
- extends Scalatest DSL with some whitespace treatment methods( """ { a: 111, b: { v: 42 } } """ should equal ("{ a: 111, b: { v: 42 } }") (after being trimmed and whitespaceNormalized) )
Provides methods for working with schemas, its comparison and alignment.
-
Schema comparison returning true/false. Ignores the order of columns
SchemaUtils.equivalentSchemas(schema1, schema2)
-
Schema comparison returning difference. Ignores the order of columns
SchemaUtils.diff(schema1, schema2)
-
Schema selector generator which provides a List of columns to be used in a select to order and positionally filter columns of a DataFrame
SchemaUtils.getDataFrameSelector(schema)
-
Dataframe alignment method using the
getDataFrameSelector
method.SchemaUtils.alignSchema(dataFrameToBeAligned, modelSchema)