Multiple contents

[scheiblr]

An implementation of multiple contents might be necessary in near future to us. We save each section of a text as one content, i.e. Attachment with its headline as title.
For better NLP results it would be important not to put the contents together prior sending it to AHD, but separately sending each one for its own.

The current implementation would no allow this as it expects to return one text blob (https://github.com/miracum/ahd2fhir/blob/master/ahd2fhir/utils/resource_handler.py#L304).

[scheiblr]

We dropped that idea after looking at it more closely, as the single contents are not referrable, but only the entire DocumentReference. Thus, we model such documents as a Bundle with an identifier containing multiple DocumentReferences with one content only.