Details
-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
None
-
False
-
-
False
Description
Deploy type
ODH Dashboard UI
Version
Red Hat OpenShift Data Science 1.33.0
Environment
OpenShift Version: 4.13.9
Pipelines Version: 1.12.1
Current Behavior
While designing a pipeline in the workbench, I added an output file in one of the node.
The output file I added is a directory path formatted as "openvino/*". This directory contains model binaries.
The pipeline fails because the * is encoded as %2A when executed in the Data Science pipeline.
According to Elyra pipelines docu, wildcards in the output path is supported. https://elyra.readthedocs.io/en/latest/user_guide/pipelines.html#output-files
Expected Behavior
The pipeline should accept wildcards in the node's output file path.
Steps To Reproduce
1. Create an elyra pipeline in the workbench.
2. Add one python file node, the python file should generate a file under a subdirectory "model"
3. Configure the node to have an Output file
4. Set the value of the output file to "model/*"
5. Export the pipeline as KFP yaml file
6. Import the KFP yaml file onto the pipeline server
7. Run the pipeline
Workaround (if any)
I had to manually declare output files for each file in the sub directory. It works if you only have few files.
----
Migrated from GitHub: https://github.com/opendatahub-io/data-science-pipelines-tekton/issues/170