-
Notifications
You must be signed in to change notification settings - Fork 23
Feature Ideas
| contents |
|---|
| ckanext-resource-currency-harvester |
| ckanext-search-results-visualization |
Develop an extension which Publishes an API for generalizing management of resource currency metadata. The extension will be composed of:
- An REST endpoint and API for registration of external resource currency metadata sources.
- a service or set of jobs that gather resource currency metadata from said sources.
- a module that uses the collected currency metadata and updates corresponding resource metadata
Mat, Colin, Greg and I met with Vivid solutions to discuss Vivid's need for additional metadata fields. There were a few fields that made sense to add others that were problematic. Mat had a stroke of insight and proposed the following:
- Vivid creates a CSV where columns correspond to the metadata fields they require.
- each row will correspond to and existing metadata record - this can be the
idof a corresponding metadata record - the additional columns will correspond to the additional metadata fields.
- This CSV will be updated by Vivid.
- Vivid will creates a metadata record for this CSV and add it as a resource.
- By adding the CSV as a resource it becomes accessible though CKAN's public APIs.
- vivid can develope dashboards against the API to query and get the information they want.
We described this solution with Leo. Leo had a more brilliant idea. PyETS already does this in some fashion, or the intention is to do so. Rather than have PyETS update metadata one way, Vivid do it another way, and hypothetically another client do it another, lets generalized the solution. We can use the solution for ourselves and other clients can also use it.
Sources of resource currency can be: simple CSVs generated by DB dumps or scripted REST Services Oracle Data REST Service
- We pull data on a configured schedule.
- We publish a Notification End point.
- Clients notify the service through the trigger end-point.
- The service then pulls the data.
The current search page is bonkers. an unauthenticated user sees a 2008 result count, some cluttered facets they have to scroll down to use and, one page of results. I want to see a visualization that gives me an immediate impression, of what is contained in those 2008 results so I can navigate and discover data quickly and successfully.