The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.

NAME

NNexus::Concepts - API for Manipulation of NNexus Concepts and Concept harvests

SYNOPSIS

  use NNexus::Concepts qw(flatten_concept_harvest diff_concept_harvests clone_concepts links_from_concept);
  $new_concepts = flatten_concept_harvest($indexed_concepts);
  ($delete_concepts,$add_concepts) = diff_concept_harvests($old_concepts,$new_concepts);
  $cloned_concepts = clone_concepts($concepts);
  @links = links_from_concept($concept);

DESCRIPTION

This package contains a range of convenience utilities for accessing and mainpulating NNexus concepts, realized via Perl hash references.

An Indexed concept is a hashref of the form:

  {
    categories => [ list of categories ],
    scheme => 'categorization scheme',
    concept => 'natural language word or phrase',
    synonyms => [ list of synonym words and phrases ],
    url => 'http://example.com'
  }

A Flattened concept is a simplified hashref of the form:

  {
    category => 'a single category',
    scheme => 'categorization scheme',
    concept => 'natural language word or phrase'
    url => 'http://example.com'
  } 

Another finer point is that when a concept is being registered, its resource/object of origin is contained in a hash key "url". When it is being registered and later retrieved from the NNexus knowledge base, that key is renamed to "link" and potentially "multilinks", when several URLs exist for the same concept. The distinction is useful to discern between the processing stages, but could also be refactored away at a later stage.

METHODS

$new_concepts = flatten_concept_harvest($indexed_concepts);

Given a list of indexed concepts with their synonyms, this routine flattens them into a simple array of concepts and performs morphological normalization on the natural language phrases, as provided by NNexus::Morphology.

($delete_concepts,$add_concepts) = diff_concept_harvests($old_concepts,$new_concepts);

Given two arrays of old and new (flattened) concepts, the routine computes their difference, returning a list of concepts to be deleted (i.e. exclusively old) and a list of concepts to be added (i.e. exclusively new).

$cloned_concepts = clone_concepts($concepts);

Clones the concept hash and returns the new hash reference. Works on an array reference of concepts, as we tend to clone in bulk.

Smart extraction of links from a concept hashref, supports both the single "link" key as well as the bulk "multilinks" field. Returns an array of URL strings.

AUTHOR

Deyan Ginev <d.ginev@jacobs-university.de>

COPYRIGHT

 Research software, produced as part of work done by 
 the KWARC group at Jacobs University Bremen.
 Released under the MIT License (MIT)