Get Document Content

Stream a document's raw PDF bytes to its owner.

Built for the field-placement editor's pdf.js viewer, which needs the sender's own upload rendered in the browser -- something no route here offered before (signers get their copy via the token-scoped GET /api/sign/{token}/document in signing.py; the owner had nothing).

Deliberately type-agnostic: every documents row this queries -- draft, template, one_off alike -- has its bytes at gcs_path, so ownership is the only thing gating this route, not document_type.