Improve prompt config UX; support JSONPath bug fix

[ohltyler]

Description

Misc improvements PR:

• improves the logic on when to expose the "Configure prompt" section in the ML processor configs. Rather than exposing by default for ML search response processors, instead only expose if there is a present 'prompt` parameters via the connector blueprint, regardless of context (ingest,search req, search resp). In general, this will require some consistency on the expected blueprints to contain such field for such use cases - ongoing discussions with ML team to sort this part out.
• supports / is consistent with the bug fix in ML ingest processors, (see here)such that we do not always default to an array obj. Updates the underlying jsonpath call from query -> value, where value returns the first hit in the target obj, instead of wrapping all hits in an arr.
• Minor wording improvements of the 'Tools' tabs
• changing hardcoded prompt field to a constant. we may have more, and/or have a regex for determining when a connector has a configurable "prompt"/"prompt template" etc. (tracked via Make prompt parameter field more flexible #393)

Check List

• Commits are signed per the DCO using --signoff

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[saimedhi]

Hello @ohltyler, would it be possible to provide a screen recording showcasing any changes from the user's perspective?