Jonny Saunders

See the commits in https://github.com/linkml/linkml/pull/1980 for trying out several versions :)

> the older generators wrap calls with redirects of stdout, and use print statements, which redirect to a string. this is confusing this has gotta get broken out to its...

idk i was just making it up, basically function neutral changes like code cleanup. we can remove and delete it np

(go ahead and delete i have no strong feelings at all and it doesn't *exactly* fit - i was applying it here bc i figured it was an uncontroversial change...

Re: perf, I doubt it, I would think that coercion would be more expensive bc the strict types fail faster, but I can profile. Re: validation vs. Instantiation, one can...

For the sake of recordkeeping - here I added the `strict` mode to the inner types of the `AnyShapeArray` for pydantic 2 - https://github.com/linkml/linkml/pull/1887/commits/0d6926a0324e9aedba19aa47bcbcf4b2c42221df core schemas are cool and transparent...

oh that's fun, we have a probabilistic test failure. It looks like the generated documentation for `NucleosomeModification.md` is nondeterministic, putting the mixins in a random order. that must have been...

``` E - ": "https:/w3id.org/linkml" E ? ^ ^ E + ": "https:\\w3id.org\\linkml" E ? ^^ ^^ E } E } ``` Go home windows youre drunk

so that looks wrong: https://github.com/sneakers-the-rat/linkml/blob/2e7fac785f59ecded70768b737be2e18a158dadb/tests/test_base/__snapshots__/types.context.jsonld#L13

Probably depends on: - https://github.com/linkml/linkml/pull/1982 - https://github.com/linkml/linkml-runtime/pull/310 to pass. i'll wait until those get merged bc i aint about to boot up a windows vm to test this lmao