Changelog

v0.1.17

Updates regarding the bibxml serializer:

  • When deciding whether to output <referencegroup> as root element, rely on presence of “includes”-type relations alone, rather than also require absence of titles.

  • Fix an issue with serializing Internet Drafts to XML

v0.1.16

  • Allow multiple BibliographicItem.version entries

v0.1.15

  • Make BibliographicItem.docid a list (and required property)

v0.1.14

  • Corrected a problem with abstract conversion (bibxml serializer)

v0.1.11

  • Fixed CodeCov badge syntax in README/PyPI description.

v0.1.10

  • bibxml serializer now takes abstract’s format into account. For abstracts with application/x-jats-xml format (typically, coming from Crossref DOI data), paragraphs are now extracted as <t> sub-elements.

v0.1.9

  • Added prefix and addition fields to PersonName model.

  • bibxml serializer now attempts to fill in fullname attribute on an <author> even if PersonName.completename is absent (formatting a string using whatever parts of the name are available).

Anchors and identifier scope

  • This starts a switch to anchor attribute values formatted based on primary identifiers, rather than using identifiers with scope property set to anchor (which will be phased out).

    So far this is only implemented for Internet Drafts (see serializers.bibxml.anchor.format_internet_draft_anchor(), which get_suitable_anchor() now delegates to if a docid with type matching “internet-draft” is detected on the item.

v0.1.8

  • Added BibliographicItem.version field and the corresponding VersionInfo class. It follows LutaML models in that draft is expected to be either a string or not defined, not an array of strings like the RNC grammar suggests. This may be subject to change.

v0.1.7

  • Fixed re-imports.

v0.1.6

  • Added re-imports for models classes in relaton.models for convenience.

  • Added changelog.