Hey experts, I am ingesting a sample MachineLearning Model into the Datahub. The Icon for ML Models at “Explore your Metadata” section is showing up and counting 1 ML Model. When I click on it I get an empty result list. When I search for the name of the Model I find it. Is this a bug or am I missing something? Any help is appreciated!

The browsepath aspect didn't generate correctly. Hence it is not showing up when you browse. Is the sample data from the project GitHub repo?

Thanks for the quick response! I adapted the GitHub Repo showcase from the bootstrap data. So I need something like this as an aspect?

{
  "com.linkedin.pegasus2avro.common.BrowsePaths": {
    "paths": [
      "/feast/test_feature_table_all_feature_dtypes"
    ]
  }
}

Actually I don't think you need to specifically ingest the browsepath aspect, the system should automatically generate its own default when an entity is created. But I'm guessing that the browsepath it generated was invalid, like /feast//xxx. The // will cause an empty folder to be shown instead. You can try docker exec I to MySQL and see what is the browsepath record for the entity, if you're interested

I just ingested the browsePath aspect with it and it worked.

hey Mike! glad your situation was taken care of here. just curious, would you be able to get the browse path for any of the entities you discussed that aren't showing up properly? I'd like to know what caused the issue in the first place so we can try to prevent that from happening again!

Hey Chris, sure I am happy to help. What do you mean with “get the browse path”, I just figured the path out on my own and then added it to the manual ingestion JSON

@dazzling-appointment-34954 So this is all working?

Hey John, yes with the additional browsepath it is working. Nevertheless maybe you can elaborate here for other people with a similar problem why for certain entities we need to specify a browsepath and for others not? That would be very helpful 🙂

Yeah I think we just don't have default browse paths implemented for the experiemental entities (all ML related entities)

@big-carpet-38439 another little question: do you know whether it is right now possible to ingest data for summary, group and features tab of ML models? I can not find any information, is this stuff still working?

It is! But totally agree we need better sample data for all of our Ml entities