Zingg on Databricks Example Error

[mechevarria]

Describe the bug
When running the Databricks Zingg example, the code fails at the line

zingg.writeLabelledOutputFromPandas(candidate_pairs_pd,args)

To Reproduce
Steps to reproduce the behavior:
An example repo with the notebook is here
This example uses Unity catalog volumes instead of deprecated dbfs paths and was run on a 15.4 LTS runtime

Expected behavior
Expected for the example to completely run

Screenshots
If applicable, add screenshots to help explain your problem.

Additional context
Add any other context about the problem here.

[sonalgoyal]

Before you execute the cell, make sure that the zingg variable has been initialized with label(cell with heading Prepare for user labeling) and not trainMatch phase. It seems that Zingg is not set to label. Can you please check and report back if you still see an issue?

[mechevarria]

@sonalgoyal looks like the notebook got a little further today rerunning all the cells. I did hit another issues a few cells down when generating docs

[sonalgoyal]

Seems like the documenter is unable to write to the volume at the docs location. Could it be a permission issue?

[mechevarria]

I'm the owner of the volume and the cluster. The example project was able to write to the other folders in the volume so I'm not sure where the permission issue is popping up from. I also manually created the docs folder and got the same problem

[mechevarria]

I'm also running into a separate issue in another environment where trainMatch is failing

[sonalgoyal]

I'm the owner of the volume and the cluster. The example project was able to write to the other folders in the volume so I'm not sure where the permission issue is popping up from. I also manually created the docs folder and got the same problem

In that case, this is likely happening since UC volumes are not supported on the Zingg Community Version.

[sonalgoyal]

I'm also running into a separate issue in another environment where trainMatch is failing

How many records have you labelled? This error indicates that either matches ornon matches are 0 in the labeled data. If you have labeled sufficiently, we have seen this issue go away when the cluster is restarted.