Christian Kadner

There is some clean-up work to be done before we can publish the MLX API Client to PyPI.

- Quickstart with Docker Compose always had delete instructions - Tommy did add delete instructions to the Kubernetes deployment doc [MLX Deployment on Kubernetes or OpenShift](https://github.com/machine-learning-exchange/mlx/blob/main/docs/mlx-setup.md#delete-the-mlx-deployment) - @Tomcli can you...

Hey Phu, glad to see you back. Please go ahead!

@Tomcli -- Given the DAX integration work with READMEs for each dataset, addressing this issue may no longer be necessary. Should we close it?

Thanks @Tomcli -- this is a good hole to plug, especially when URL go out of date. I added an issue to help keep URL working: https://github.com/machine-learning-exchange/katalog/issues/43 However, we also...

This issue was [commented](https://github.com/machine-learning-exchange/mlx/pull/228#discussion_r715948750_) on by @Tomcli on PR #228 > One of the use cases we want to cover here is when the `readme_url` is not valid. We want...

Alternative ways to address this issue: 1. During **upload**, the **API** can validate the `readme_url` -- This is not perfect since that URL may no longer be available after some...

Thanks Andrew. I will take a look at this

@drewbutlerbb4 -- I cannot reproduce this bug. How did you enable the upload in the Quickstart with Docker Compose since that was not working anymore (#97)?

Hi @drewbutlerbb4 -- I have been trying to reproduce this without luck. Could this have been a issue with concurrent requests or maybe related to caching?