Creating an algorithm

All algorithms inherit from JIPipeAlgorithm. An algorithm consists of following parts:

JIPipe comes with different base algorithms that provide different feature sets:

Algorithm type Purpose
JIPipeAlgorithm The base class of all algorithms. It provides no included functionality outside of absolutely necessary ones.
JIPipeParameterSlotAlgorithm This algorithm allows users to run multiple parameter sets by optionally enabling an additional slot Parameters.
JIPipeSimpleIteratingAlgorithm A JIPipeParameterSlotAlgorithm that has one input slot and iterates over the input rows.
JIPipeIteratingAlgorithm A JIPipeParameterSlotAlgorithm that can have multiple input slots and match annotations of the input data to create data batches.
JIPipeMergingAlgorithm A JIPipeParameterSlotAlgorithm that can have multiple input slots and match annotations of the input data to create data batches. The difference to JIPipeIteratingAlgorithm is that a batch can have duplicate items per slot.
JIPipeSingleIterationAlgorithm A JIPipeParameterSlotAlgorithm that can have multiple inputs and acts like a JIPipeMergingAlgorithm where all data rows are always merged into one batch.

Any algorithm should have the following basic structure:

// Annotates documentation to the algorithm
@JIPipeDocumentation(name = "My Algorithm", description = "Does something")

// Sets the algorithm category
@JIPipeNode(nodeTypeCategory = MiscellaneousNodeTypeCategory.class)

// Input and output slots autoCreate automatically creates the slots if set to true and no slot configuration was provided
@JIPipeInputSlot(value = ImagePlusData.class, slotName = "Input", autoCreate = true)
@JIPipeOutputSlot(value = ImagePlusData.class, slotName = "Output", autoCreate = true)

// You can add multiple JIPipeCitation annotations to provide citations for this node only
@JIPipeCitation("Additional citation")

public class MyAlgorithm extends JIPipeAlgorithm {

    This is the main constructor of the algorithm.
    It contains a reference to the algorithm info that contains
    some important metadata
    public MyAlgorithm(JIPipeNodeInfo info) {

    A deep copy constructor. It is required.
    Please do not forget to deep-copy all important fields
    public MyAlgorithm(MyAlgorithm original) {
        // Deep-copy additional fields here

    The workload is run in this functon.
    You can also have the ability to inform the user about the current algorithm status
    for more complex algorithms.
    You can also query if the user requested cancellation
    public run(JIPipeProgressInfo progress) {
        // Run your workload here

To register the algorithm and provide it with an Id, use JIPipeJavaExtension:

@Plugin(type = JIPipeJavaExtension.class)
public class MyExtension extends JIPipeDefaultJavaExtension {

    // ... See previous tutorial for other methods
    public void register(JIPipe jiPipe, Context context, JIPipeProgressInfo progressInfo) {
        // Registers our algorithm with a unique ID and an icon
        registerNodeType("my-algorithm", MyAlgorithm.class, UIUtils.getIconURLFromResources("actions/viewimage.png"));