{% extends "layout.html" %} {% import "macros/profile.html" as profile %} {% block content %} {# The namespace/prefix definitions used to shorten IRIs #} {% set prefixes, error = from_file(conf.template_path+'/prefixes.json').fetch_tree() %}
This section attempts at reconstructing the profiled shapes for each class. The profile is deduced based on structural patterns for a selected class and their descriptive statistics aggregated by distinct properties.
The structural pattern is a triple pattern of the type (subject,predicate,object), which selects subjects that are instances of a given class and on that collection of individuals identifies all unique predicates used, then for each of those predicate identifies the type of the values used with it (i.e. either the class or the data type). The pattern can be summarised as this: (ClassOfSubjects, distinctPredicate, ClassOfObject). The descriptive statistics account for (a) the number of unique subjects, (b) the number of unique objects (c) number of total occurrences of the pattern (d) minimal, (e) maximal and (f) average number of occurences.
Each class states its precision: exact when every instance was read, or first N of M when only the first N instances the store returned were read. Absent values are shown as —.