DataHub Querier

Description

It is a javascript class created to make it easier and more effective to search for knowledge graphs among those made available by DataHub.

The site uses a CKAN (API) through which it was possible to automatically create a JSON file containing all the KGs and Datasets maintained by the site.

Therefore it is possible to analyze and query this file, in particular you can perform a brutal search and a search by tag ang multiple tag.

You can also choose to rank the results in 4 different ways:

• name (default)
• size
• authority
• centrality

Basic Usage

Install with npm: npm install datahub-querier

// First of all you have to require the package in the code
var dh_querier = require('datahub-querier');

// ..then you have to initialize the lc_querier
var querier = new dh_querier();

Now you are ready to exploit all the functions:

BRUTAL SEARCH

var results = querier.brutalSearch('keyword', 'rankingMode'); 
//rankingMode(optional) is one of['name', 'size', 'authority', 'centrality']


TAG SEARCH

var results = querier.tagSearch('keyword', 'tag', 'rankingMode'); 
//choose one of the tags from DataHub json structure.


MULTITAG SEARCH

var results = querier.multiTagSearch('keyword', 'tag_1', 'tag_2', 'tag_3', ...,  rankingMode);
// you perform the query on several tags.

Available methods for a datahub-querier instance

• brutalSearch(target) : For each dataset in DataHub, it searches within all tags for the regular expression containing the target.
• tagSearch(target, tag): For each dataset, it searches within the specified tag for the regular expression containing the target.
• multiTagSearch(target, ...tags): For each dataset, it searches within the specified tags the regular expression containing the target.
• filterResults(result, ...tags): It's a filter to return in the resulting JSON only tags specified.
• generalSorting(result, mode): It's a dispatcher method to execute the ranking algorithm specified in mode parameter.
• sortResultsBySize(results): Sorts results by triples number.
• sortResultsByName(results): Sorts results in alphabetic order using the name.
• sortResultsByAuthority(results): Sorts results by authority using the pagerank algorithm.
• sortResultsByCentrality(results): Sorts results by centrality using the centrality algorithm.