get metadata from DOIs

[jonaslenz]

• Metadata attached to single DOIs can be accessed from different registration agencies (RA)
• the RA, which is behind a particular DOI, can be retrieved through https://doi.org/ra/, e.g. https://doi.org/ra/10.1594/GFZ.TR32.2 shows that the Seibert Database is registered at DataCite
• Each RA has its own API to access metadata, e.g. DataCite, [CrossRef](curl -LH "Accept: application/json" http://dx.doi.org/10.1038/nrd842)

DataCite should be in our first scope, as they are focused on depositing content (e.g. datasets). A multiple RA query is (in an ancient commit) implemented in Zotero - maybe we can also learn from them.

[jonaslenz]

There are several pypackages for the zenodo API - does any of us suits our (meta)data retrieval?

• [ ] https://pypi.org/project/pyzenodo3/#history
• [ ] https://pypi.org/project/pynodo/ <-- seems currently broken
• [ ] https://github.com/lgloege/zenodopy <-- seems currently broken
• [ ] https://pypi.org/project/zenodo-rest/ <-- is deprecated

[jonaslenz]

A first prototype query with caching of retrieved metadata is implemented in the webapp for: doi.org(query RA) -> Datacite (get metadata) -> Zenodo data files (if Dataset is on Zenodo)

@jandevaty: Shall we implement this workflow as functions of the metadata generation class?

[jandevaty]

yes, we should. In general we need to separate the data/metadata/files/xml/whatever manipulation and (any) front end actions. The front-end app should not directly access any data/metadata/files providers - it should always use the back-end objects to do so, because the back-end needs to be completely separate to be usable as an API

[jonaslenz]

Outside Zenodo download URLs may not be retrievable through an API. e.g. Roger dataset: doi: 10.6094/UNIFR/151460 API of Freiburg (without download URL): https://freidok.uni-freiburg.de/jsonApi/v1/publications?doi=10.6094/UNIFR/151460

[jonaslenz]

seems to be solved allready