Create a _development snapshot_ release of sssom schema with pydantic instead of regular data classes

[matentzn]

As per @cmungall using the pydantic data classes could significantly improve the performance of the sssom tooling. I would like to try the following:

• Switch to pydantic data classes on a branch
• Make a candidate release (ensure this creates a development snapshot on pypi, not a regular one)
• Create a branch in sssom py using that
• See what breaks (report here)
• If all ok, create a candidate release of sssom-py
• Import into oak, see what breaks, report here
• Make proper releases

[hrshdhgd]

Now the data classes we currently have are source from a YAML file and then Linkml's python generator creates the data class file. Your vision is to replace all that by 1 pydantic data class file, correct? Does this mean cleaning out linkml dependencies?

[matentzn]

Thanks for this question.. I was hoping to be able to pass instances of the pydantic data classes to the linkml converters the same way I was able to pass the old data classes?