settings.py is too confusing about bibxml-service HOT 2 CLOSED

Settings are minimally documented for a maintainer already aware how things work generally. (Note that configuration is not intended to be edited by non-maintainers.)

But yes, documentation is very insufficient, and this certainly needs to be addressed (#44, #45).

To avoid duplicate effort, extensive documentation will be added after we update architecture & configuration model to accommodate new approach to structuring sources, which became obviously required to accommodate mixed source scenarios such as I-Ds/Datatracker or xml2rfc datasets that don’t map to a relaton-/bibxml-data repositories, and the new docid-centric model for citation query.

In meantime, a quick additional summary for how settings work right now (pre abovementioned update), if that helps:

• DATASET_SOURCE_OVERRIDES — configures dataset source, only if it’s different than default.
• AUTHORITATIVE_DATASETS — which datasets are authoritative (by the criteria you have provided, authoritative datasets are considered i-ds, rfcs and rfcsubseries)
• EXTERNAL_DATASETS — which of the known datasets are external sources (not indexable, like DOI and later Datatracker)
• KNOWN_DATASETS — which dataset IDs do we have at all
• LEGACY_DATASETS — dataset IDs for xml2rfc compatibility APIs. Its documentation spans almost 100 lines, so I doubt it’s hard to tell how this one is used.

from bibxml-service.

The latest update changes settings:

• now we have RELATON_DATASETS, EXTERNAL_DATASETS and AUTHORITATIVE_DATASETS
• documentation now references the settings

from bibxml-service.