API inconsistency: MgetDoc uses "id" while hits use "doc_id"

[davidkelley]

Problem\n- MgetDoc serializes the document identifier as "id" while search hits use "doc_id". This is user-visible and inconsistent across APIs.\n\nImpact\n- Client code needs special-casing between search and mget responses.\n\nPotential solutions\n- Breaking change: rename MgetDoc field to "doc_id" in responses.\n- Transitional: add a new "doc_id" field while keeping "id" for backwards compatibility (deprecate "id"), or keep output as "doc_id" and add serde alias for input if needed.\n- Alternatively align on "_id" across APIs (would also be breaking).\n\nNotes\n- Update HTTP/core tests accordingly if/when changed.\n