Report absolute HPS for specs that model it

[DIzosimov]

Third of three, split out of #1787 following review. Independent of the other two.

Noting up front: the review raised a preference not to emphasise overall HPS numbers, since the model produces one but in-game results vary a lot. That's a fair product call and this PR is the one to reject or trim if you'd rather not surface it — the other two don't depend on it. Posting it because there's a real bug underneath regardless (see below).

The bug that exists today

The competitive alternatives list prints a hardScore delta labelled "HPS". hardScore is an intellect-equivalent ranking number, not healing, so that label is simply wrong. It predates this change. Even if the HPS display isn't wanted, that label should be corrected.

What this adds

Sets evaluated through a cast model or a ramp sim now carry setHPS, a genuine healing figure from the same evaluation that ranks them. It's shown on the stat panel with an upgrade percentage over currently-equipped gear, and per alternative alongside the healing and percentage given up. The percentage is derived from the HPS figures themselves, so the two can't disagree.

The stat weight path deliberately reports nothing rather than a fabricated number. baseHPS there is a hardcoded placeholder, and that path pre-applies raid buffs which the cast models also apply internally, so anything derived from it would be wrong. Those specs keep the existing relative display.

Design notes

• The equipped set is evaluated through the same evalSet as the candidates, so the two figures are comparable rather than one being modelled and the other a baseline constant.
• It runs outside the ranking loop and can never influence which set wins.
• It's wrapped so a malformed equipped set costs the upgrade percentage, not the whole run.
• A dead-on 0% — which happens when Top Gear is run without adding any candidate items, so the best set is simply the player's current gear — now says so instead of reading like the comparison failed.
• The percentage renders negative correctly, for the case where the built set is worse than what's equipped.

Testing

41 suites / 204 tests passing, production build clean. The suite asserts the figure is throughput rather than the ranking score, that alternatives never out-heal the chosen set, and that the derived percentage is consistent with the raw healing gap.