yank
yank copied to clipboard
Create public Schemas for Cerberus
In follow up to #823, we want to make a YAML(s) which have the human-readable Cerberus schema's for the validators.
There are a few steps to making this happen:
- Make the
YANKCerberusValidator
accept either the currentExperimentBuilder
orExperimentBulder._db
as an initialization argument, so it has access to the instance - Take all validator/coerce/normalize/etc. functions and convert them to methods of the
YANKCerberusValidatorClass
- Design function names non-abiguously so when we write out the YAML dictionary, its all human readable
- Ensure we don't get a circular import scheme between
experiment.py
andschema
module. - Figure out how to generate these schemas as part of the docs creation because several depend on reading functions through
generate_function_schema
that I would rather not hard code - Make sure the schemas can be generated without instancing anything (pure class and static methods). This step strongly depends on the 2nd and 3rd step of this process.