Set a file field through its field manager, not the raw mutator

[ramonski]

Description of the issue/feature this PR addresses

A NamedFile or NamedImage field on a Dexterity type that also carries a BBB set<Name> mutator is written through the mutator, which stores the value exactly as it arrives. Over the API a file arrives as base64 text, so the field ends up holding the text rather than the file, and the object then fails its own validation on that field.

The data manager already goes through the field manager for the managers that normalize their value (UID references, durations, decimals). NamedFileFieldManager does precisely that: it base64-decodes the payload and builds the NamedFile or NamedImage the field is declared to hold, taking the filename and content type from the payload. It belongs in that tuple, and NamedImageFieldManager comes with it as a subclass.

Current behavior before PR

POST /senaite/@@API/senaite/v1/update
{"uid": "<laboratory uid>",
 "accreditation_body_logo": "<base64>",
 "filename": "seal.svg"}

HTTP/1.1 400 Bad Request
{"message": "{\"accreditation_body_logo\": \"wrong type\"}", ...}

The raw base64 string is written onto the field by setAccreditationBodyLogo, and the following full-object validation refuses it.

Desired behavior after PR is merged

The value goes through NamedFileFieldManager, which decodes it and stores a NamedBlobImage. The file is then served as usual:

GET /senaite/setup/laboratory/@@download/accreditation_body_logo/seal.svg
HTTP/1.1 200 OK
Content-Type: image/svg+xml

Verification

• Set accreditation_body_logo on a Laboratory over the API with a base64 payload and a filename, and fetch it back through @@download and @@images.
• A file field on a type without a BBB mutator (which already worked) is unchanged: it reached the field manager through the no-setter fallback before and reaches it one branch earlier now.