Transform

abstract class Transform

A Transform that processes intermediary build artifacts.

For each added transform, a new task is created. The action of adding a transform takes care of handling dependencies between the tasks. This is done based on what the transform processes. The output of the transform becomes consumable by other transforms and these tasks get automatically linked together.

The Transform indicates what it applies to (content, scope) and what it generates (content).

A transform receives input as a collection TransformInput, which is composed of JarInputs and DirectoryInputs. Both provide information about the Scopes and ContentTypes associated with their particular content.

The output is handled by TransformOutputProvider which allows creating new self-contained content, each associated with their own Scopes and Content Types. The content handled by TransformInput/Output is managed by the transform system, and their location is not configurable.

It is best practice to write into as many outputs as Jar/Folder Inputs have been received by the transform. Combining all the inputs into a single output prevents downstream transform from processing limited scopes.

While it's possible to differentiate different Content Types by file extension, it's not possible to do so for Scopes. Therefore if a transform request a Scope but the only available Output contains more than the requested Scope, the build will fail.
If a transform request a single content type but the only available content includes more than the requested type, the input file/folder will contain all the files of all the types, but the transform should only read, process and output the type(s) it requested.

Additionally, a transform can indicate secondary inputs/outputs. These are not handled by upstream or downstream transforms, and are not restricted by type handled by transform. They can be anything.

It's up to each transform to manage where these files are, and to make sure that these files are generated before the transform is called. This is done through additional parameters when register the transform.

These secondary inputs/outputs allow a transform to read but not process any content. This can be achieved by having getScopes() return an empty list and use getReferencedScopes() to indicate what to read instead.

Summary

Public constructors

A Transform that processes intermediary build artifacts.

Public methods

open Boolean
applyToVariant(@NonNull variant: VariantInfo)

Whether this transform should be applied to a given variant.

abstract MutableSet<QualifiedContent.ContentType!>

Returns the type(s) of data that is consumed by the Transform.

abstract String

Returns the unique name of the transform.

open MutableSet<QualifiedContent.ContentType!>

Returns the type(s) of data that is generated by the Transform.

open MutableMap<String!, Any!>

Returns a map of non-file input parameters using a unique identifier as the map key.

open MutableSet<in QualifiedContent.Scope!>

Returns the referenced scope(s) for the Transform.

abstract MutableSet<in QualifiedContent.Scope!>

Returns the scope(s) of the Transform.

open MutableCollection<File!>

Returns a list of additional (out of streams) directory(ies) that this Transform creates.

open MutableCollection<File!>

Returns a list of additional file(s) that this Transform needs to run.

open MutableCollection<File!>

Returns a list of additional (out of streams) file(s) that this Transform creates.

open MutableCollection<SecondaryFile!>

Returns a list of additional file(s) that this Transform needs to run.

open Boolean

Returns if this transform's outputs should be cached.

abstract Boolean

Returns whether the Transform can perform incremental work.

open Unit
setOutputDirectory(@NonNull directory: Property<Directory!>)

For Transforms that produce an output available though the [BuildArtifactsHolder] provider based interfaces should indicate the output directory of the produced artifact

open Unit

open Unit
transform(@NonNull context: Context, @NonNull inputs: MutableCollection<TransformInput!>, @NonNull referencedInputs: MutableCollection<TransformInput!>, @Nullable outputProvider: TransformOutputProvider?, isIncremental: Boolean)

open Unit
transform(@NonNull transformInvocation: TransformInvocation)

Executes the Transform.

Public constructors

<init>

Transform()

A Transform that processes intermediary build artifacts.

For each added transform, a new task is created. The action of adding a transform takes care of handling dependencies between the tasks. This is done based on what the transform processes. The output of the transform becomes consumable by other transforms and these tasks get automatically linked together.

The Transform indicates what it applies to (content, scope) and what it generates (content).

A transform receives input as a collection TransformInput, which is composed of JarInputs and DirectoryInputs. Both provide information about the Scopes and ContentTypes associated with their particular content.

The output is handled by TransformOutputProvider which allows creating new self-contained content, each associated with their own Scopes and Content Types. The content handled by TransformInput/Output is managed by the transform system, and their location is not configurable.

It is best practice to write into as many outputs as Jar/Folder Inputs have been received by the transform. Combining all the inputs into a single output prevents downstream transform from processing limited scopes.

While it's possible to differentiate different Content Types by file extension, it's not possible to do so for Scopes. Therefore if a transform request a Scope but the only available Output contains more than the requested Scope, the build will fail.
If a transform request a single content type but the only available content includes more than the requested type, the input file/folder will contain all the files of all the types, but the transform should only read, process and output the type(s) it requested.

Additionally, a transform can indicate secondary inputs/outputs. These are not handled by upstream or downstream transforms, and are not restricted by type handled by transform. They can be anything.

It's up to each transform to manage where these files are, and to make sure that these files are generated before the transform is called. This is done through additional parameters when register the transform.

These secondary inputs/outputs allow a transform to read but not process any content. This can be achieved by having getScopes() return an empty list and use getReferencedScopes() to indicate what to read instead.

Public methods

applyToVariant

Added in API level 3.4
@Incubating open fun applyToVariant(@NonNull variant: VariantInfo): Boolean

Whether this transform should be applied to a given variant.

Parameters
variant VariantInfo: information about the current variant.
Return
Boolean: true if the transform should be applied to a given variant, false otherwise.

getInputTypes

@NonNull abstract fun getInputTypes(): MutableSet<QualifiedContent.ContentType!>

Returns the type(s) of data that is consumed by the Transform. This may be more than one type. This must be of type QualifiedContent.DefaultContentType

getName

@NonNull abstract fun getName(): String

Returns the unique name of the transform.

This is associated with the type of work that the transform does. It does not have to be unique per variant.

getOutputTypes

@NonNull open fun getOutputTypes(): MutableSet<QualifiedContent.ContentType!>

Returns the type(s) of data that is generated by the Transform. This may be more than one type.

The default implementation returns getInputTypes().

This must be of type QualifiedContent.DefaultContentType

getParameterInputs

@NonNull open fun getParameterInputs(): MutableMap<String!, Any!>

Returns a map of non-file input parameters using a unique identifier as the map key.

Changes to values returned in this map will trigger a new execution of the Transform even if the content inputs haven't been touched.

Changes to these values force a non incremental execution.

The default implementation returns an empty Map.

getReferencedScopes

@NonNull open fun getReferencedScopes(): MutableSet<in QualifiedContent.Scope!>

Returns the referenced scope(s) for the Transform. These scopes are not consumed by the Transform. They are provided as inputs, but are still available as inputs for other Transforms to consume.

The default implementation returns an empty Set.

getScopes

@NonNull abstract fun getScopes(): MutableSet<in QualifiedContent.Scope!>

Returns the scope(s) of the Transform. This indicates which scopes the transform consumes.

getSecondaryDirectoryOutputs

@NonNull open fun getSecondaryDirectoryOutputs(): MutableCollection<File!>

Returns a list of additional (out of streams) directory(ies) that this Transform creates.

These File instances can only represent directories. For files, use getSecondaryFileOutputs()

Changes to directories returned in this list will trigger a new execution of the Transform even if the qualified-content inputs haven't been touched.

Changes to these output directories force a non incremental execution.

The default implementation returns an empty collection.

getSecondaryFileInputs

@NonNull open fun getSecondaryFileInputs(): MutableCollection<File!>

Deprecated: replaced by getSecondaryFiles()

Returns a list of additional file(s) that this Transform needs to run. Preferably, use getSecondaryFiles() API which allow eah secondary file to indicate if changes can be handled incrementally or not. This API will treat all additional file change as a non incremental event.

Changes to files returned in this list will trigger a new execution of the Transform even if the qualified-content inputs haven't been touched.

Any changes to these files will trigger a non incremental execution.

The default implementation returns an empty collection.

getSecondaryFileOutputs

@NonNull open fun getSecondaryFileOutputs(): MutableCollection<File!>

Returns a list of additional (out of streams) file(s) that this Transform creates.

These File instances can only represent files, not directories. For directories, use getSecondaryDirectoryOutputs()

Changes to files returned in this list will trigger a new execution of the Transform even if the qualified-content inputs haven't been touched.

Changes to these output files force a non incremental execution.

The default implementation returns an empty collection.

getSecondaryFiles

@NonNull open fun getSecondaryFiles(): MutableCollection<SecondaryFile!>

Returns a list of additional file(s) that this Transform needs to run.

Changes to files returned in this list will trigger a new execution of the Transform even if the qualified-content inputs haven't been touched.

Each secondary input has the ability to be declared as necessitating a non incremental execution in case of change. This Transform can therefore declare which secondary file changes it supports in incremental mode.

The default implementation returns an empty collection.

isCacheable

open fun isCacheable(): Boolean

Returns if this transform's outputs should be cached. Please read Javadoc if you would like to make your transform cacheable.

isIncremental

abstract fun isIncremental(): Boolean

Returns whether the Transform can perform incremental work.

If it does, then the TransformInput may contain a list of changed/removed/added files, unless something else triggers a non incremental run.

setOutputDirectory

open fun setOutputDirectory(@NonNull directory: Property<Directory!>): Unit

For Transforms that produce an output available though the [BuildArtifactsHolder] provider based interfaces should indicate the output directory of the produced artifact

Parameters
directory Property<Directory!>:

setOutputFile

open fun setOutputFile(@NonNull file: Property<RegularFile!>): Unit

transform

open fun transform(
    @NonNull context: Context,
    @NonNull inputs: MutableCollection<TransformInput!>,
    @NonNull referencedInputs: MutableCollection<TransformInput!>,
    @Nullable outputProvider: TransformOutputProvider?,
    isIncremental: Boolean
): Unit

Deprecated: replaced by transform(TransformInvocation).

transform

open fun transform(@NonNull transformInvocation: TransformInvocation): Unit

Executes the Transform.

The inputs are packaged as an instance of TransformInvocation

  • The inputs collection of TransformInput. These are the inputs that are consumed by this Transform. A transformed version of these inputs must be written into the output. What is received is controlled through getInputTypes(), and getScopes().
  • The referencedInputs collection of TransformInput. This is for reference only and should be not be transformed. What is received is controlled through getReferencedScopes().
A transform that does not want to consume anything but instead just wants to see the content of some inputs should return an empty set in getScopes(), and what it wants to see in getReferencedScopes().

Even though a transform's Transform#isIncremental() returns true, this method may be receive false in isIncremental. This can be due to

  • a change in secondary files (getSecondaryFiles(), getSecondaryFileOutputs(), getSecondaryDirectoryOutputs())
  • a change to a non file input (getParameterInputs())
  • an unexpected change to the output files/directories. This should not happen unless tasks are improperly configured and clobber each other's output.
  • a file deletion that the transform mechanism could not match to a previous input. This should not happen in most case, except in some cases where dependencies have changed.
In such an event, when isIncremental is false, the inputs will not have any incremental change information:
Parameters
transformInvocation TransformInvocation: the invocation object containing the transform inputs.
Exceptions
IOException if an IO error occurs.
InterruptedException
TransformException Generic exception encapsulating the cause.