Schemas cleanup following model ADR #3

[Rotorsoft]

Link to Issue

Closes: #8826

Description of Changes

• Follows schema conventions in ADR

Test Plan

• Basic testing since there are no major changes here, keeping an eye on date parsing

[timolegros]

I disagree with this #8824 (comment). Unless I am mistaken most if not all date types modified in this PR are directly related to Postgres TIMESTAMP columns. As such I think it may be beneficial to have 2 predefined date types with documentation:

// Use for any date that is required in the database but is automatically set by Sequelize or Postgres
`PG_REQUIRED_DATE = z.coerce.date().optional()`

// Use for any date that can be `null` in the database
`PG_OPTIONAL_DATE = z.coerce.date().nullish()`

Naming can of course be changed the idea is simply that we have 2 groups of dates and always remembering to use coerce and nullish or optional will get messy.

[Rotorsoft]

I disagree with this #8824 (comment). Unless I am mistaken most if not all date types modified in this PR are directly related to Postgres TIMESTAMP columns. As such I think it may be beneficial to have 2 predefined date types with documentation:

// Use for any date that is required in the database but is automatically set by Sequelize or Postgres
`PG_REQUIRED_DATE = z.coerce.date().optional()`

// Use for any date that can be `null` in the database
`PG_OPTIONAL_DATE = z.coerce.date().nullish()`

Naming can of course be changed the idea is simply that we have 2 groups of dates and always remembering to use coerce and nullish or optional will get messy.

We can do this, but I'm questioning the value add of these utils...all dates should be coerced and the optional vs nullish decision is by column anyway... just trying to avoid unnecessary utilities, but if we decide to go with this, I would do a PG_DATE = z.coerce.date() only and leave the optional vs nullish to the schema

[timolegros]

We can do this, but I'm questioning the value add of these utils...all dates should be coerced and the optional vs nullish decision is by column anyway... just trying to avoid unnecessary utilities, but if we decide to go with this, I would do a PG_DATE = z.coerce.date() only and leave the optional vs nullish to the schema

Let's discuss with the others in office hours cause for me without adding nullish or optional to the utils then every time I add a date I need to figure out whether I need to use nullish or optional (and what each means) instead of just using the REQUIRED or OPTIONAL util which makes this decision easy.