elyra
elyra copied to clipboard
elyra-ai
Elyra extends JupyterLab with an AI centric approach.
**Is your feature request related to a problem? Please describe.** Currently export produces a file with the pipeline name, e.g. `my-pipeline.py` or `my-pipeline.yaml` for Kubeflow Pipelines environments. With the newly...
**Describe the issue** With the updates to the build in #2447 the `generateClassName` parameter of `StylesProvider` is no longer being run consistently. This is causing some css classes to not...
Hi, I'm using the pipeline executor locally, and I will appreciate the possibility of executing only a subset of nodes of a pipeline. Sometimes it happens that you build a...
**Is your feature request related to a problem? Please describe.** With the upcoming release of notebook 7 alongside jupyterlab 4, lab extensions will now be automatically supported in notebook. We...
**Is your feature request related to a problem? Please describe.** No. We want to add functionality to Elyra and a Model Explorer would be a nice low hanging fruit **Describe...
**Describe the issue** While investigating potential options how to support partial pipeline execution, I've noticed that nodes that are implemented using _generic components_ are not [cached by Kubeflow Pipelines](https://www.kubeflow.org/docs/components/pipelines/overview/caching/). This...
With the merge of https://github.com/jupyterlab/jupyterlab/pull/12365 the Code Viewer is now a part of core Jupyterlab on master. Once this feature is released, most likely in 4.0, we should switch to...
**Is your feature request related to a problem? Please describe.** We're working with a client where we deal with large amount of data, even during development. For some parts we...
**Is your feature request related to a problem? Please describe.** Based on kubeflow pipelines documentation https://www.kubeflow.org/docs/components/pipelines/sdk/connect-api/, one can use kfp client to invoke the other workflow, but needs to poll...
When performing large workloads, it does not make sense to use node inputs/outputs to propagate data as it would be very cost-intensive. One idea would be to map an S3/COS...
