How we combined CKAN with WordPress for opendata.swiss

Transcript

1. How we combined CKAN
   with WordPress
   Stefan Oderbolz
   CKANCon 2016
   4th October 2016
   

   View Slide

2. View Slide

3. Requirements
   ● Use CKAN for the data catalogue
   ○ DCAT-AP Switzerland metadata standard
   ○ Multilingual content
   ● Arbitrary pages (FAQ, About, Blog)
   ○ Multilingual content
   ● Manageable forms (Contact, surveys etc.)
   ● Harvesting of data in DCAT-AP Switzerland format
   ● ...
   

   View Slide

4. Solution: let’s combine CKAN with WordPress
   ● Loosly coupled: they call each others APIs
   ● We use Apache as a proxy
   ○ WordPress runs on port 80
   ○ CKAN runs on port 8080
   ○ Specific routes are proxied on WordPress
   ● User management only on WordPress
   

   View Slide

5. View Slide

6. View Slide

7. Goal: one administration interface
   ● Our WordPress plugin (soon Open Source!) let’s us manage CKAN entities:
   ○ Datasets
   ○ Groups
   ○ Organizations
   ○ Harvester
   ● We save all data in WordPress and make API calls to “sync” them to CKAN
   ● Form has some usability issues
   ○ Too long, no possibility to collapse
   ○ Save button is on the top
   

   View Slide

8. Intermezzo: What is DCAT-AP?
   ● DCAT = Data Catalog vocabulary
   ○ Vocabulary to describe catalogs, dataset and distributions
   ● AP = Application profile
   ○ Ruleset of mandatory, optional or conditional fields
   ● E.g. “DCAT-AP for data portals in Europe” (DCAT-AP EU) is a standard for
   public sector data in Europe
   ● DCAT-AP Switzerland is compatible to DCAT-AP EU, but has some additional
   fields or fields that are currently not used
   More Information: http://handbook.opendata.swiss/en/library/ch-dcat-ap
   

   View Slide

9. View Slide

10. View Slide

11. DCAT-AP XML Import/Export
    ● Easy transfer from test system to production
    ● Reference implementation for data owners, transition to DCAT-AP XML
    harvester
    

    View Slide

12. Harvester Dashboard
    ● Manage harvesters from
    WordPress
    ● Start, stop & clear harvester
    jobs
    ● Indicator of status
    ● Check the error log of
    previous jobs
    

    View Slide

13. View Slide

14. View Slide

15. APIs are key
    ● To exchange data between the two systems we use APIs on both sides
    ● CKAN
    ○ Standard API calls for CRUD operations on all entities
    ○ Custom API calls where needed (e.g. dataset count total + per group, to have only one call)
    ● WordPress
    ○ Expose WordPress entities as custom API (e.g. “Application”)
    ○ To increase performance we save some API responses for 10 minutes in Redis (e.g. Dataset
    count on frontpage)
    

    View Slide

16. @odi
    Questions?
    

    View Slide