-
Notifications
You must be signed in to change notification settings - Fork 0
Description
Let's create an extensible way to process outputs of batch-processed modules.
Currently, module outputs will be collected and shown as a Table after batch processing:
batch-processor/src/main/java/org/scijava/batch/ModuleBatchProcessor.java
Lines 116 to 119 in 6052ea3
| for (Entry<String, Object> output : outputs.entrySet()) { | |
| outputTable.set(output.getKey(), outputTable.getRowCount() - 1, | |
| output.getValue()); | |
| } |
The outputFolder parameter is currently unused, but was foreseen to be used to save image outputs for each iteration when processing modules that output images.
This output processing should be provided in an extensible way by BatchOutputProcessors, so that we'd have an ImageWriterBatchOutputProcessor (likely in a different repository) that takes care of collecting outputs of a compatible type (Dataset, Img, ...) and saves them to the provided directory.
In this case, the ModuleBatchProcessor needs to be made aware of the required additional inputs (in this case, File outputFolder) and we need to dynamically add inputs to the dialog depending on the output processor that is available.
This needs a bit more thought, and I'd be happy about feedback from others.