Conversation
| __tablename__ = "validations" | ||
|
|
||
| id = Column(Integer, primary_key=True, index=True) | ||
| entity_type = Column(String, nullable=False, index=True) |
There was a problem hiding this comment.
I am not sure I understand :( . Can you give an example of 1 instance of this ? and how it would be consumed ?
There was a problem hiding this comment.
It might be easier to look at the schema rather than the data model. So let's say the user contributes a new morphology. The validation logic looks up the morphology entity type in the table then applies the "must_pass_to_upload" validations. If these are successful it runs the validations in "must_run_upon_upload" which annotates the entity, and "must_pass_to_simulate" which generates additional validations needed for the artifact to be simulated.
There was a problem hiding this comment.
ok, then it is not a validation, it is rather something like a "validation plan".
also what does the list of string contains ? I know it is a list of validations. but how do we go from a string to the actual execution of a code ?
There was a problem hiding this comment.
This is the validation plan schema stored in the database. The string contains a relative path in a github repo, for example "/morphologies/validate_morphology.py" I wrote a simple "validate_morphology.py" script that sees if it can be read in by NeuroM and returns an error if not. There will need to be some support logic that processes the lists, which I can add to entitycore once we are satisfied with this plan.
2 participants
I'm opening this to PR to get feedback on the validation schema.