Derive observation type from its category and fix cagemate matching

[labkey-martyp]

Rationale

Two independent fixes, to observation entry and to cagemate matching. The Observations form let a user pick any observation type but stored every entry as Clinical, so behavior observations recorded there were filed as clinical and dropped out of the behavior views; the form cannot know the right value up front because it depends on which type the user picks for each row, so the type is now derived on save from that type's category. The behavior forms had the same mismatch from the other direction: their Daily Observations shortcut bypassed the type-filtered picker and wrote clinical-category observations labeled as behavior. Separately, the cagemates query treated a housing row with no cage as though it were a location, so every animal whose location had not resolved was reported as a cagemate of every other.

Rows already saved with a mismatched observation type need a one-time data fix; this change only affects new entries.

Related Pull Requests

None.

Changes

• The Observations form no longer defaults an observation's type. The trigger script derives it from the selected observation type's category, while every other form continues to set the type explicitly, including scheduled entries that inherit it from their order.

• The Daily Observations shortcut is now opt-in per form section rather than always present, so it appears only on the clinical forms.

• Cagemate matching now keys off active housing rows with a known cage.