fix(mimic-iv): do not treat CANCELLED cultures as positive

[acnimma]

Summary

• Exclude org_itemid 90760 (CANCELLED) from positiveculture in suspicion_of_infection.sql
• Also exclude org_name = 'CANCELLED'
• Update postgres and duckdb dialect copies

Test plan

• Source + dialect copies exclude 90760 / CANCELLED
• Specimen with only CANCELLED → positive_culture = 0
• Specimen with a real organism → positive_culture = 1

positiveculture treated any non-NULL, non-empty org_name as positive except NEGATIVE (90856). CANCELLED cultures (org_itemid 90760) are not microbial growth, but were counted as positive and could inflate positive_culture on the suspicion-of-infection table. Exclude CANCELLED the same way NEGATIVE is excluded.

Fixes #1677

[Chessing234]

yeah treating CANCELLED as growth was pretty misleading for soi. excluding 90760 and the org_name string both seems fine (belt and suspenders). only thing i'd double-check is whether any other org_itemids are similarly non-organism (contaminated/etc) — not blocking, just curious.

[Chessing234]

excluding cancelled cultures is overdue. worth noting in a comment next to the filter so future edits don't reintroduce them.

[Chessing234]

good catch — cancelled cultures aren't growth. covering itemid 90760 and org_name 'CANCELLED' in all three dialects is the right shape. maybe add a one-line comment near the CASE that 90760 is the cancelled organism itemid so the next reader doesn't have to dig microbiologyevents for it.