Python Graph.preferredLabel方法代码示例

# 需要导入模块: from rdflib.graph import Graph [as 别名]
# 或者: from rdflib.graph.Graph import preferredLabel [as 别名]
def write_mappings():

    # Include all pairs:
    # ll = pair_association(unique_pairs, loglike)

    # Only pairs with frequency >= 5:
    unique_pars_5 = np.array([a.split('!') for a, b in pc.items() if b >= 5])

    ll = pair_association(unique_pars_5, loglike)

    print 'Write mappings'
    lls = sorted(ll.items(), key=lambda x: x[1])

    ww = np.array(ll.values())
    mn = ww.mean()

    print "- Mean LL is {:2f}".format(mn)
    print "- {:.2f} % is >= mean LL".format(float(ww[ww >= mn].shape[0]) / ww.shape[0])
    print "- {:.2f} % is < mean LL".format(float(ww[ww < mn].shape[0]) / ww.shape[0])

    # Whether to lookup DDC labels and add them to the mapping sheet
    addDdcLabels = False

    if addDdcLabels:
        # Load WebDewey data
        g = Graph()
        for x in glob('../../webdewey/DDK23/*.ttl'):
            print x
            g.load(x, format='turtle')

    fsj = re.compile('.*\(Form\)')
    with open('mappings.csv', 'w') as f:
        writer = csv.writer(f, delimiter='\t')
        for x in lls[::-1]:
            if x[1] < mn:
                break

            q = x[0].split('!', 1)

            if fsj.match(q[0]):  # Utelat form
                continue

            if addDdcLabels:
                lab = g.preferredLabel(URIRef('http://dewey.info/class/' + q[1] + '/e23/'), labelProperties=[SKOS.prefLabel, SKOS.altLabel])
                if len(lab) != 0:
                    lab = lab[0][1].value
                else:
                    lab = '(no label)'
                # Term, Dewey, Dewey Caption, Loglike
                writer.writerow([q[0], q[1], lab.encode('utf-8'), x[1]])
            else:
                # Term, Dewey, Loglike
                writer.writerow([q[0], q[1], x[1]])

# 需要导入模块: from rdflib.graph import Graph [as 别名]
# 或者: from rdflib.graph.Graph import preferredLabel [as 别名]
    def load(self, filename):
        """
        Note: This loader only loads categories and mappings
        """
        graph = Graph()
        graph.load(filename, format=self.extFromFilename(filename))

        logger.info('Read %d triples from %s', len(graph), filename)

        skosify.infer.skos_symmetric_mappings(graph, related=False)

        # Load mappings
        n_mappings = 0
        n_memberships = 0
        for tr in graph.triples_choices((None, [SKOS.exactMatch, SKOS.closeMatch, SKOS.broadMatch, SKOS.narrowMatch, SKOS.relatedMatch], None)):
            source_concept = tr[0]
            res_id = self.vocabulary.id_from_uri(source_concept)
            if res_id is not None:
                shortName = str(tr[1]).split('#')[1]
                try:
                    self.vocabulary.resources[res_id].add('mappings.%s' % shortName, str(tr[2]))
                    n_mappings += 1
                except KeyError:
                    logger.warning('Concept not found: %s', res_id)

        # Load categories
        for tr in graph.triples((None, RDF.type, UOC.Category)):
            cat_lab = graph.preferredLabel(tr[0], lang='nb')[0][1].value
            cat_id = '' + tr[0]

            cat = Concept().set_type('Category')
            cat.set('id', cat_id)
            cat.set('prefLabel.nb', Label(cat_lab))
            self.vocabulary.resources.load([cat])


            for tr2 in graph.triples((tr[0], SKOS.member, None)):
                uri = str(tr2[2])
                res_id = self.vocabulary.id_from_uri(uri)
                if res_id is not None:
                    try:
                        self.vocabulary.resources[res_id].add('memberOf', cat_id)
                        n_memberships += 1
                    except KeyError:
                        logger.warning('Concept not found: %s', res_id)

        # Load number of ccmapper mapping candidates
        for tr in graph.triples((None, LOCAL.ccmapperCandidates, None)):
            source_concept = tr[0]
            res_id = self.vocabulary.id_from_uri(source_concept)
            if res_id is not None:
                shortName = str(tr[1]).split('#')[1]
                try:
                    self.vocabulary.resources[res_id].set('ccmapperCandidates', int(tr[2]))
                except KeyError:
                    logger.warning('Concept not found: %s', res_id)

        # Load ccmapper mapping state
        for tr in graph.triples((None, LOCAL.ccmapperState, None)):
            source_concept = tr[0]
            res_id = self.vocabulary.id_from_uri(source_concept)
            if res_id is not None:
                shortName = str(tr[1]).split('#')[1]
                try:
                    self.vocabulary.resources[res_id].set('ccmapperState', tr[2])
                except KeyError:
                    logger.warning('Concept not found: %s', res_id)

        logger.info('Loaded %d mappings and %d category memberships from %s', n_mappings, n_memberships, filename)