CKAN harvester endpoint

This is a small HTTP endpoint that gathers datasets that are publishable to
opendata.swiss, transforms them to be compatible with the CKAN harvester and
outputs a format (a rigid form of RDF/XML) that the harvester can read.

The format expected by the harvester is described on
https://handbook.opendata.swiss/fr/content/glossar/bibliothek/dcat-ap-ch.html (fr/de).

In order to be considered as a "publishable" dataset by this endpoint, a
dataset must follow the following conditions:

- it has the dcat:Dataset type
- it has one and only one dcterms:identifier
- it has a dcterms:creator
- it has schema:workExample with value
- it has schema:creativeWorkStatus with value
- it does not have schema:validThrough
- it does not have schema:expires

The endpoint copies the following properties with their original value.
Make sure they follow the CKAN spec.

- dcterms:title
- dcterms:description
- dcterms:issued
- dcterms:modified
- dcat:contactPoint
- dcat:theme
- dcterms:language
- dcat:landingPage
- dcterms:spatial
- dcterms:coverage
- dcterms:temporal
- dcterms:keyword (literals without a language are ignored)

The following properties are populated by the endpoint:

- dcterms:identifier

If the original dcterms:identifer already contains an "@", it is copied
as-is. Otherwise, an identifier is created with the shape
@, where "creator slug" is
the last segment of the URI of the value of dcterms:creator.

- dcterms:publisher

The original dcterms:publisher value is used as rdfs:label of the
final dcterms:publisher.

- dcterms:relation

Populated from dcterms:license.

TODO: define valid values for license

- dcterms:accrualPeriodicity

Supports both DC (http://purl.org/cld/freq/) and EU
(http://publications.europa.eu/ontology/euvoc#Frequency) frequencies.
DC frequencies are transformed into EU ones.

- dcat:distribution

Populated from schema:workExample.
Only takes work examples with a schema:encodingFormat into account.
Each distribution is built in the following way, from the properties of the work example:

- dcterms:issued is copied as-is
- dcat:mediaType populated from schema:encodingFormat
- dcat:accessURL populated from schema:url
- dcterms:title populated from schema:name
- dcterms:rights populated from schema:identifier of the dataset's dcterms:rights
- dcterms:format populated from schema:encodingFormat, with the following mapping:
- text/html -> HTML
- application/sparql-query -> SERVICE
- other -> UNKNOWN

Usage

This should be used as a Trifid plugin.

The following options are supported:

- endpointUrl: URL to the SPARQL endpoint
- user: User to connect to the SPARQL endpoint
- password: Password to connect to the SPARQL endpoint

Configuring Trifid to use @zazuko/trifid-plugin-ckan is easy, just add the following in your configuration file:

``yaml
plugins:
# …other plugins

ckan:
module: "@zazuko/trifid-plugin-ckan"
paths: /ckan
config:
endpointUrl: https://some-custom-endpoint/
# user: root
# password: super-secret
`

and update the config fields with correct informations.

Do not forget to add it to your Node dependencies:

`sh
npm install @zazuko/trifid-plugin-ckan
`

With this configuration, the service will be exposed at /ckan and will require the organization query parameter, like this: /ckan?organization=…`.

This will trigger the download of a XML file.