search

neontribe / Linked_Development

Linked Development
1 stars 1 forks source link

Inconsistencies in existing data #44

Closed tobybatch closed 11 years ago

tobybatch commented 11 years ago

If we look at this URL

http://api.ids.ac.uk/openapi/eldis/get_all/documents/full

In some sections the data held does not seem to fully represent the data presented. e.g:

        "category_theme_ids": [
            "C38", 
            "C80", 
            "C1521", 
            "C2", 
            "C38"
        ],

Seems to indicate 5 themes. Yet in the category_theme_array section of the results there are only three themes enumerated. What's all that about then?

practicalparticipation commented 11 years ago

In this case it appears C2 is the parent object for all themes, and C38 is being duplicated, and is present because it is the parent of C80 and C1521.

So - this is one of the cases, and around the detailed theme objects in general, where we may end up diverging slightly from the exact responses returned by the IDS API.

We might need to investigate which features existing tools and widgets are relying upon - as generating the hierarchy information to output in API responses could be costly for us to do in SPARQL queries.

In the first instance I would suggest we focus on getting category_theme_array with just the directly attached categories (via dcterms:subject) and then we can later look at fetching the skos:conceptScheme relationships that a category is embedded within.