Clone a WEB file's S3 folder content from a source file into this
(target) file. The standard URL-download + extract pipeline can't
handle WEB content because it's a folder of files, not a single zip.

Body — exactly one of:

• {"source_file_id": "<source_id>"} — server-side S3-to-S3 copy
  using the source File row (must be in the same DB / same region).
• {"signed_urls": [{"url": ..., "target_relative_path": ...}, ...]}
  — HTTP-download each signed URL into the target's S3 path.
  Used for cross-DB / cross-region promotions.

Async — returns 202 immediately. The target file's status transitions
from PENDING to VIEWABLE (or FAILED) once the background copy finishes.