Skip to content

JetBrains-Research/npy

Repository files navigation

JetBrains Research Tests Status Maven Central

npy

npy allows to read and write files in NPY npy and [NPZ] npy formats on the JVM.

Installation

The latest version of npy is available on [Maven Central] maven-central. If you're using Gradle just add the following to your build.gradle:

repositories {
    mavenCentral()
}

dependencies {
    compile 'org.jetbrains.bio:npy:0.3.5'
}

With Maven, specify the following in your pom.xml:

<dependency>
  <groupId>org.jetbrains.bio</groupId>
  <artifactId>npy</artifactId>
  <version>0.3.5</version>
</dependency>

The previous versions were published on Bintray. They can be downloaded from GitHub Releases.

Examples

NPY

val values = intArrayOf(1, 2, 3, 4, 5, 6)
val path = Paths.get("sample.npy")
NpyFile.write(path, values, shape = intArrayOf(2, 3))

println(NpyFile.read(path))
// => NpyArray{data=[1, 2, 3, 4, 5, 6], shape=[2, 3]}

NPZ

val values1 = intArrayOf(1, 2, 3, 4, 5, 6)
val values2 = booleanArrayOf(true, false)
val path = Paths.get("sample.npz")

NpzFile.write(path).use {
    it.write("xs", values1, shape = intArrayOf(2, 3))
    it.write("mask", values2)
}

NpzFile.read(path).use {
    println(it.introspect())
    // => [NpzEntry{name=xs, type=int, shape=[2, 3]},
    //     NpzEntry{name=mask, type=boolean, shape=[2]}]

    println("xs   = ${it["xs"]}")
    println("mask = ${it["mask"]}")
    // => xs   = NpyArray{data=[1, 2, 3, 4, 5, 6], shape=[2, 3]}
    //    mask = NpyArray{data=[true, false], shape=[2]}
}

Limitations

The implementation is rather minimal at the moment. Specifically it does not support the following types:

  • unsigned integral types (treated as signed),
  • bit field,
  • complex,
  • object,
  • Unicode
  • void*
  • intersections aka types for structured arrays.

Building from source

The build process is as simple as

$ ./gradlew jar

Note: don't use ./gradlew assemble, since it includes the signing of the artifacts and will fail if the correct credentials are not provided.

Testing

No extra configuration is required for running the tests from Gradle

$ ./gradlew test

However, some tests require Python and NumPy to run and will be skipped unless you have these.