Record the Publication DOI

[sneumann]

Hi,
while we probably thought about this ourselves, this came up in NFDI4Chem and also regarding our JSON-LD metadata:
Our PUBLICATION information is human interpretable, but we're not keeping a persistent identifier to the publication in our records. Like a DOI. Changing this probably goes together with #194 and more structure in our tags as mentioned in #157 (comment)

This was also raised by Alisdair (of course :-) ) in

... we are working with OpenAIRE so that Bioschemas can be used as a mechanism to populate the OpenAIRE research graph. For this case, and for other consumers, it would be helpful if you could provide the DOI as an IRI value for the citation property. This could be done in two ways:

1. Leave your existing string version and add a second citation property that has the doi as the value
2. Model the text of the citation as a ScholarlyArticle with the DOI given with the @id property

Both approaches should work for people consuming the markup and allow them to link your data with existing data that they have.

Yours,
Steffen