make type slot *not* required

[wdduncan]

In order for change sheets to work correctly, the entity being updated needs to have value in its type slot. So, we need to require the type slot.

@dehays Do you thinktype needs to be required for all class in the schema? I think "yes", but we may want to discuss more.

For reference, this issue came about as a result of #88

cc @cmungall @turbomam @sujaypatil96

[wdduncan]

Update:
In the latest changes, the type slot in the schema has been deprecated. So, the type slot should not be required.
See this thread for discussion concerning deprecation of type slot: microbiomedata/nmdc-schema#248

In fact, I have some code in change sheets where I'll need to un-require the type.

[ssarrafan]

Moving this to March per @wdduncan as he plans to close before the end of the week

Note from Bill:
HI Set. Here is an update:
microbiomedata/nmdc-schema#134
This is an ongoing issue that will need to be passed on to Mark or Sujay (not sure which).
#89
I updated the comment on this. I should be able to get the change sheet edit done before I leave.
microbiomedata/nmdc-schema#195
I am working with Sujay on this. I should be able to close before the week’s end.
#46
This is an ongoing issue that will need to be passed on to Mark or Sujay (not sure which).
microbiomedata/nmdc-server#555
This is a lot of conversation in this thread. But, it looks like Mark has taken this one over.

[ssarrafan]

@turbomam @dehays @cmungall this was an old issue still assigned to Bill. I re-assigned it to @turbomam. Should this issue be closed or moved to a future sprint?

[wdduncan]

Unfortunately, I was not able to fix this issue before leaving. In the file metadata.py here is the line that needs to be fixed:

nmdc-runtime/nmdc_runtime/api/core/metadata.py

Line 171 in f4f51e9

raise Exception("Cannot determine the type of class for ", id_)

It may suffice to simply set class_name = "" when a type could not be determined. However, this needs to be tested. There are couple of testing notebooks I set up for running tests:

changesheets-testing.ipynb
changesheets-example.ipynb

You may need to enlist @dehays and @dwinston for help explaining how changesheets work.

[ssarrafan]

Thanks @wdduncan for the update from Florida!
@turbomam wants to continue this to April so I'll move it to the April sprint

[turbomam]

I feel like it should be super straightforward to set type's required attribute to false, but @wdduncan's comment about changing a line in the autogenerated metadata.py makes me think I'm missing the point. Also, I certainly don't have a good sense of all of the ways an edit like this could affect all of the systems that consume that are guided by the nmdc-schema. Like parts of the portal? Or some ingest steps?

type has never been are required slot for biosamples in the DH templates.

Also, we have been recently talking about "overloading" the biosample class by instantiating it for material entities in a sample processing chain. Does that touch on this issue at all?

[dwinston]

Taking a quick look at this, it seems that collection_name is guaranteed to be set at the point of the changesheet processing code that seeks to set class_name, and these two values are one-to-one and connected in the json schema, e.g. the 'biosample_set' collection is known to be a list of objects of type 'Biosample'. So, I think type can be deduced unambiguously.

[wdduncan]

The type slot is now deprecated:
https://microbiomedata.github.io/nmdc-schema/attribute_value_type/#slot-type-deprecated

So, I think the goal was to remove any need for it in the code base.

[dwinston]

fixed via 54439bf