[AI.a2ml] Add validation-chain-over-tuple-match idiomatic pattern #49

Proposal

Add an (idiomatic-pattern) for Validation chain over tuple match. case (a, b, c, d, e, f, g) of (Ok, Ok, Ok, ...) => ... blows ambiguity-resolution budgets at N≥5. The Validation applicative chain is the right shape.

Source observation

ProvenBridge's extractLevelData had a 7-tuple case match ((devicesR, zonesR, guardsR, ztR, hasPbxR, pbxIpR, pbxWxR)). Idris2's default %ambiguity_depth of 3 couldn't resolve it; even at 12 the cascade into surrounding code was fragile. Splitting into 4-tuple + 3-tuple matches fixed it. The Validation applicative gives this as one-liner MkRecord <$> a <*> b <*> c <*> ... <*> g.

Suggested patch (LESSONS.md L5)

(pattern
  (name "validation-chain-over-tuple-match")
  (composition "Validation applicative + <*>")
  (purpose "Accumulate errors across N independent extractions without N-tuple matches.")
  (anti-pattern "case (a, b, c, d, e, f, g) of (Ok, Ok, Ok, ...) — N>=5 tuple matches blow type inference.")
  (sketch "MkRecord <$> a <*> b <*> c <*> d <*> e <*> f <*> g"))

Filed from idaptik Wave 3 pilot. See LESSONS.md and PILOT.md. Idaptik commit hyperpolymath/idaptik@aef38db.