Repository navigation
Parse a Submission's species_type at the read (null drafts, legacy values) #435
Description
Split from #425.
Species.species_typeisstring(src/species/types.ts:12), thoughSpeciesTypeexists.Species.is_cares_speciesisnumber(src/species/types.ts:17); should beboolean.NewSpecies.pointClass?: number | nulland the update input (src/species/curation.ts:45,:92) should usePointClass.
Each DB read that builds a Species has to parse at the boundary (isSpeciesType / admitPointClass), not cast. Same for the species_type: string copies in db/submissions.ts, db/activity.ts, db/members.ts, db/cares.ts.
Activity
Type a Species' type, Point class and CARES flag, parsed at the read (#…
jra3 commentedon Oct 28, 2026
Note
🤖 Claude Opus 5.5 responding on behalf of John Allen
#592 did most of this, so I'm narrowing the issue to what remains:
Species.species_typeisSpeciesType,is_cares_speciesisboolean, andbase_pointsisPointClass | null. Every read that builds a Species parses withspeciesOfRow.NewSpecies.pointClassandSpeciesUpdate.pointClasstakePointClass.db/members.tsanddb/cares.tsare gone.Still open:
species_type: stringonSubmission(src/db/submissions.ts:39) andSubmissionApprovedData(src/db/activity.ts:34). These can't be parsed the same way yet.submissions.species_typeis nullable with no CHECK, and drafts are saved from a partial form. We need to decide first:species_typeparses to:SpeciesType | null, or a separate draft type.SpeciesType: a migration that normalizes them plus a CHECK constraint, or refusing at the read. This needs a count of the distinct values in prod first.Relabelled
ready-for-humanuntil those are settled.[-]Tighten Species public types: SpeciesType, boolean CARES flag, PointClass[/-][+]Parse a Submission's species_type at the read (null drafts, legacy values)[/+]