[Dev UI] Evaluators - default eval for deep equal

[DenisVCode]

Overview

Genkit provides number of already created evaluators. I think deep equal evaluator should be one of them for comparing outputs of structured responses.

Related Discord thread: https://discord.com/channels/1255578482214305893/1315783915964993730/1319736165800218734

I will probably create custom evaluator because I don't want to wait for this, so when I am done I would be happy to send a PR if that is welcome.

User goal(s)

This is useful for using evaluation to check if upgrading model or changing prompt did not break previously known good generations, for cases like text extraction, where the output should always be the same.

Requirements

Acceptance Criteria

• 1. Deep equal of reference and structured output
• 2. Returns pass/fail

Nice to have:

• Diff display