difi/dcat-harvester

Rest api for datasett

Closed this issue · 8 comments

hoyum commented

Api som leverar datasett med alle felt i json-format. API må kunne ta mot unik nøkkel som parameter for uthenting av eit spesifikt datasett. Ref guid frå API RSS.

What kind of requests do you intend to be included as part of the API?

Primærbehovet er å kunne hente alle data knytt til eit datasett.Feks. tittel, beskrivelse, format osv. i eit strukturert format.

Dette skal nyttast for enkelt å kunne bygge detaljvisningen av eit datasett.

APIet må då kunne ta mot ein parameter, feks den unike nøkkelen kvart datasett har i RDF grafen (@id).

Vidare bør APIet levere ut noko som dette:

{"dataset": {
"title":"",
"description":"",
"distribution": [
{"title":"","description":"","format":""},
{"title":"","description":"","format":""},
{"title":"","description":"","format":""},
],
"xxxx":"xxx",
}
}

hoyum commented

Kan henge saman med #23

Apart from a basic JSON dump request, how else do you expect/intend for end users to consume this service? Assuming this REST endpoint is to be used for search with Elasticsearch as the backend, here are some basic ideas to start with:

  • Aggregations (facets) on particular fields
  • Filters on particular fields - within and across data sets to which the user has access
  • Generic search with definable query and result fields

Elasticsearch is obviously capable of quite a lot of complicated aggregations so we also need to very clearly define what kinds of analyses and aggregations to allow.

hoyum commented

I think these scenarios are the most relevant to begin with.

hoyum commented

Set this task on hold.

I think we get the required API directly from ElasticSearch

Now that data is in Elaticsearch and we're keeping Elasticsearch open for queries, can probably close this issue.

hoyum commented

Dekka av Elasticsearch