Add a migration to remove possible old values in organization table

[Macroz]

Add a migration to remove possible old values in organization table. Consider making the validation give a warning, but the code to ignore any unknown values. Optionally validate on save and/or disregard unknown values.

update organization set data = data - 'organization/modifier' - 'organization/last-modified';

Related to #2823

clojure.lang.ExceptionInfo: Value cannot be coerced to match schema: {:organization/modifier disallowed-key, :organization/last-modified disallowed-key}
	at schema.coerce$fn__3627$coercer_BANG___3632$fn__3633$fn__3634.invoke(coerce.clj:52)
	at clojure.core$mapv$fn__8535.invoke(core.clj:6979)
	at clojure.lang.PersistentVector.reduce(PersistentVector.java:343)
	at clojure.core$reduce.invokeStatic(core.clj:6885)
	at clojure.core$mapv.invokeStatic(core.clj:6970)
	at clojure.core$mapv.invoke(core.clj:6970)
	at rems.db.organizations$get_organizations_raw.invokeStatic(organizations.clj:42)
	at rems.db.organizations$get_organizations_raw.invoke(organizations.clj:39)
	at rems.db.organizations$get_all_organization_roles.invokeStatic(organizations.clj:83)
	at rems.db.organizations$get_all_organization_roles.invoke(organizations.clj:81)
	at rems.middleware$wrap_context$fn__35620.invoke(middleware.clj:73)
	at rems.middleware$wrap_user$fn__35615$fn__35616.invoke(middleware.clj:65)
	at rems.logging$with_mdc_STAR_.invokeStatic(logging.clj:21)
	at rems.logging$with_mdc_STAR_.invoke(logging.clj:9)
	at rems.middleware$wrap_user$fn__35615.invoke(middleware.clj:64)
	at ring.middleware.anti_forgery$wrap_anti_forgery$fn__670.invoke(anti_forgery.clj:94)
	at rems.middleware$wrap_api_key_or_csrf_token$fn__35612.invoke(middleware.clj:57)
	at buddy.auth.middleware$wrap_authentication$fn__31678.invoke(middleware.clj:60)
	at rems.auth.auth$wrap_uses_valid_api_key$fn__32220.invoke(auth.clj:40)```

[Macroz]

In the migration we dropped the columns, but some of the values have snuck up in the JSON blobs. Now that there exists a validation for the data, the extra values prevent normal use.