feat: improve linkml error messages for regex validation

[daniel-ji]

Currently when there is a pydantic regex validation error (for example, a pattern is not matched for an ontology id, the error message does not include the correct pattern / ontology for which the value should match. This error message is now improved, so that users (and LLM agents) know how to fix their patttern.

For example, previously, an error message would have looked like this:

FAIL: "/home/daniel/cryoet-data-portal-backend/ingestion_tools/dataset_configs/10444.yaml":
        - Value error, Value error, Invalid id format: CVC_0045: datasets.0.metadata.cell_strain.id

But now, it looks like this:

FAIL: "/home/daniel/cryoet-data-portal-backend/ingestion_tools/dataset_configs/10444.yaml":
        - Value error, Value error, Invalid id format: CVC_0045. Valid formats include: ['WBStrain[0-9]{8}$', '^NCBITaxon:[0-9]+$', '^CVCL_[A-Z0-9]{4,}$', '^CC-[0-9]{4}$']. Field description: The ontology identifier for the cell strain.: datasets.0.metadata.cell_strain.id

Also a small fix to the template.yaml file.

[daniel-ji]

how to actually get the patterns that are used for a given field: provide the linkml metadata for that field; and then search the global linkml metadata that contains all types (including their regex patterns for the custom types we have built for the ontologies)

[daniel-ji]

returns a decorator that wraps the class so that when regex validation errors happen; their error messages are caught and wrapped with more details and then thrown again.