Open Mbrownshoes opened 7 years ago
Analysis please Jared?
Producing RDF files from catalogue works. When adding the dcat ext to the configuration file, it has to be listed before the bcgov ext because of routing order for getting the RDF graph of a dataset. eg.
ckan.plugins = ...
dcat dcat_json_interface ...
edc_rss edc_dataset edc_app edc_geo edc_ngeo edc_webservice
...
To use the harvester feature of the dcat ext, it requires ckanext-harvest and redis to be installed in the env first.
I had some issues when trying to use the dcat's processor on the cli. I was using the demo ckan site's output and piping it to the consume mode of the processor to write the dataset mapping to a json file, but it failed on missing translator object.
There are custom profiles that can be made for customizing how the values defined in the RDF graph are mapped to CKAN and viceversa. There's more info on writing custom profiles as well in the dcat ext readme.
This is also relevant to the Google Dataset Discovery initiative. https://developers.google.com/search/docs/data-types/dataset. https://ai.googleblog.com/2017/01/facilitating-discovery-of-public.html
Google recommends using JSON-LD rather than RDF or DCAT, but it supports all 3.
Currently BCDC datasets are not findable by Google Dataset Search https://toolbox.google.com/datasetsearch
We'd like to implement the dcat extension so that our content can be shared with other data catalogues, such as http://open.canada.ca/
https://github.com/ckan/ckanext-dcat