Closed Viqsi closed 1 year ago
I propose an alternative. Instead of writing yet another complex handler for arbitrary expressions, instead require that a pedigree cohort be defined by a construct variable that has as one of it's categoric values 'affected' (or 'include', whatever you guys like so long as it is The Rule), and a minimum count of 'affected' individuals for the family to be included.
The concern there is that making construct variable creation a hard prerequisite for use of the ascertain pedigree feature is troublesome at best. I think a case can be made for restricting it to categorical variables only (so if someone wants to use something non-categorical, they have to downcode it first), but making it required for ALL uses is IMO putting too many roadblocks in the user's way.
I could live with a single specified value of a categoric variable that is applied to a cohort defining the included families. But if it is complicated in any way, those complications should be expressed in construct variables.
On Thu, Apr 6, 2023 at 7:08 PM Jo Valentine-Cooper @.***> wrote:
The concern there is that making construct variable creation a hard prerequisite for use of the ascertain pedigree feature is troublesome at best. I think a case can be made for restricting it to categorical variables only (so if someone wants to use something non-categorical, they have to downcode it first), but making it required for ALL uses is IMO putting too many roadblocks in the user's way.
— Reply to this email directly, view it on GitHub https://github.com/MathematicalMedicine/diverweb-issues/issues/134#issuecomment-1499733262, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA52M5A63FDJK4DUXC6ISOTW75EFTANCNFSM6AAAAAAWNTW6YQ . You are receiving this because you were assigned.Message ID: @.***>
That's probably fair.
Veronica agrees with a categoric value being the inclusion/exclusion criteria as of 4/10/'23 status meeting.
This was taken care of in MathematicalMedicine/diverRPC@92330c3
Spec document is attached: Ascertain Peds Module Design v1.docx
Basically, what's needed is a schema for passing requested parameters from diverweb to the backend. That comes under diverRPC, so we arguably both get dragged into it.
I'm planning on putting together a first draft in diverRPC, though.