Given a web annotation, guess the URL of its actual subject.
I wish I knew a better word for it. In a discussion about an image, a reply to my comment will have my comment as its target, but it's most likely about the image, too. In that case, the subject is the image.
Since there are so many things we could discuss using web annotations, I need a term that works for all of it – images, media files, physical books, dinosaur bones, people, living animals, ideas about future projects.
To make things even harder, I'm not a native speaker of American English, so I hope my choice of "subject" is good enough.
This module exports one function that holds some methods:
anno is a plain old JS object that conforms to the
W3 annotation-model.
Returns the URL as a string, or throw an error (see .foundNone() below)
if no subject was found.
Throw an error complaining that no subject target was found.
The error will have .type === 'AnnoNoSubjectTarget'.
Like guessSubjectTargetUrl(), but when no subject was found,
don't throw an error, rather just return false.
(In versions up to 1.0.2, this was described in README as the default behavior, but it was never actually implemented.)
anno is as above, but multi-target annotations are supported.
Returns an array of URLs. The array will be either empty,
or contain truthy values.
For well-formed supported input, that "truthy" more specifically means
means non-empty strings, which should be URLs. Which may however be
relative to the annotation's URL, so we can't always know for sure.
The array will also hold some methods:
.primary()will return the first item if there is one, orfalse..additional()will return an array with all but the first item, orfalseif that array would be empty.- Treating the first target as "primary" is not a violation of Anno Model section 3.2.6 "Cardinality of Bodies and Targets" because it requires equal relatedness only for bodies.
:TODO:
- Needs more/better tests and docs.
ISC