Kotlin PreviewGenerator (PrevGen)

Kotlin PreviewGenerator is a lightweight compiler plugin writ ten with Kotlin Symbol Processing (KSP). Preview Generator creates preview instances for data classes, which can be used in Jetpack Compose view previews.

Navigation

• Setup
• Usage
• Examples
• Roadmap
• Feedback and contact us
• License
• Contribute
• Configurations

Setup

1. Add the KSP plugin in the main build.gradle:

   plugins {
       // ...
       id("com.google.devtools.ksp") version "1.9.21-1.0.15"
   }

2. Add the Sonatype Maven URL:

   repositories {
       // ...
       mavenCentral()
   }

3. Add KSP and the module into Gradle dependencies:

   dependencies {
       implementation("io.github.vram-voskanyan.kmp:PreviewGenerator:1.0.0") // take latest from Maven central
       ksp("io.github.vram-voskanyan.kmp:PreviewGenerator:1.0.0")
   }

4. Sync the project, and you are good to go.

Usage

1. Add a @DataPreview annotation to the class that needs to have a generated preview.
2. Rebuild the project. It will generate a Preview class inside build\ksp\[BUILD_VARIANT]\ClassName.kt.

Examples

@DataPreview
data class DummyClass(val name: String, val age: Int, val loginDate: Long)

val previewValue = dummyClassPreview  

// dummyClassPreview is a generated class which is: DummyClass(name = "Ryan", age = 85, loginDate = 1705600601029)
// Note: `loginDate` is a Timestamp ;)

Do Not Do This

These are examples that will not work:

    class A(items: List<A>)

In this case, it will attempt to create an endless items.

    data class B(enumItem: EnumValue)

Currently, enums and sealed classes are not supported. Make the type nullable in this case:

    data class B(enumItem: EnumValue?)

For more configurations and setups, please refer here: Configurations.md

Example for Jetpack Compose

Medium Demo

Ongoing and Future Works

Since this project was developed only during non-office hours, not everything is covered. For example, the following types are not currently supported: Non-Data classes, Enum, Map, List (in progress).

We also have some exciting features in mind for implementation, depending on community interest:

1. Customized Data Size: Introduce an option to edit the generated data size for checking how the view behaves with large or small datasets. This may involve adding a parameter option to annotations.

2. Flexible Data Generation: Currently, the generator relies on an argument name-checker to pick values close to the parameter type. For the next step, we aim to provide an option to pass values from outside, allowing customization such as non-English values.

3. Adding caching functionality: Actually, the script is already fast, but there are plans to implement caching for pregenerated data. ;)

Contact Author

• Email: vram.arm@gmail.com
• LinkedIn: Vram Voskanyan

License

Apache License 2.0

Contributing

If you would like to contribute code you can do so through GitHub by forking the repository and sending a pull request. When submitting code, please make every effort to follow existing conventions and style in order to keep the code as readable as possible.