Develop API

[robons]

All URIs should adhere to the URL standards for GOV.UK.

API technical and data standards (v2 - 2019)

Catalogue Endpoints

• [ ] Endpoint to return a list of all RDF graphs uploaded to the repos (perhaps searchable?)
  • requires paging
• [ ] Endpoint to return a list of the historic versions uploaded for a particular RDF graph
  • requires paging
• [ ] Endpoint to upload a versioned CSV-W to the repository (An RDF version of this should be automatically generated on upload).
• [ ] Endpoint to upload a versioned RDF file to the repository.

N.B. When uploading, nothing should be returned in the 'live' listing endpoints until the whole ingestion/indexing process has been completed.

Content Acquisition Endpoints

• [ ] Endpoint to download the RDF (in ttl format) for a given version (0.1) of a graph (some-graph-identifier)
  • http://some-base-url/graphs/some-graph-identifier/0.1 with no Accept header OR Accept: text/plain (until we develop a suitable human-friendly web page) OR Accept: text/ttl
  • http://some-base-url/graphs/some-graph-identifier/0.1/ttl with any or no Accept header
  • http://some-base-url/graphs/some-graph-identifier/0.1/rdf with any or no Accept header
  • We could add more formats of RDF in the future?
• [ ] Endpoint to download the CSV (if available) for a given version (0.1) of a graph (some-graph-deitnfier)
  • http://some-base-url/graphs/some-graph-identifier/0.1 with Accept: text/csv
  • http://some-base-url/graphs/some-graph-identifier/0.1/csv with any or no Accept header
• [ ] Endpoint to download the CSV-W metadata JSON (if available) for a given version (0.1) of a graph (some-graph-deitnfier)
  • http://some-base-url/graphs/some-graph-identifier/0.1 with Accept: application/csvm+json
  • http://some-base-url/graphs/some-graph-identifier/0.1/csvw with any or no Accept header
• [ ] The above content acquisition endpoints should accept latest as a version, e.g. http://some-base-url/graphs/some-graph-identifier/latest. This should return the data associated with the latest version uploaded.
• [ ] http://some-base-url/graphs/some-graph-identifier should perma-redirect to http://some-base-url/graphs/some-graph-identifier/latest

Do we want to support the memento protocol?

https://github.com/ONSdigital/dp/blob/main/standards/API_STANDARDS.md

[canwaf]

Do we want to automatically generate a CSV-W from a given RDF? (Though clearly not something we're building now.)

[robons]

Do we want to automatically generate a CSV-W from a given RDF? (Though clearly not something we're building now.)

I don't think so. There's no way to know we'd be making a sensible CSV-W. Which records do we decide are the ones from the RDF which should be present in the CSV and which ones need to go in the metadata JSON? How do you 'correctly' decide?