1 # -*- coding: utf-8 -*-

     2 '''

     3 Created on Feb 22, 2012

     4 

     5 @author: raphv

     6 '''

     7 

     8 from django.core.management.base import NoArgsCommand

     9 from django.core.management.color import no_style

    10 from hdabo.utils import show_progress

    11 from hdabo.models import Tag

    12 from hdalab.models import Country, GeoInclusion

    13 from SPARQLWrapper import SPARQLWrapper, JSON

    14 import re

    15 

    16 class Command(NoArgsCommand):

    17     def handle_noargs(self, **options):

    18         self.style = no_style()

    19 

    20         qs = Tag.objects.filter(category__label="Localisation").exclude(dbpedia_uri = None)

    21         total = qs.count()

    22         

    23         endpoint = SPARQLWrapper("http://dbpedia.org/sparql")

    24         endpoint.setReturnFormat(JSON)

    25         sparqltext = """

    26             SELECT ?resource WHERE {

    27              { <%s> ?resource <http://dbpedia.org/ontology/Country> . }

    28             UNION

    29              { <%s> <http://dbpedia.org/ontology/country> ?resource . }

    30             }

    31         """

    32         resourceprefix = "http://dbpedia.org/resource/"

    33         identityuri = "http://www.w3.org/1999/02/22-rdf-syntax-ns#type"

    34         

    35         writer = None

    36         

    37         for i,tag in enumerate(qs):

    38             endpoint.setQuery(sparqltext % (tag.dbpedia_uri, tag.dbpedia_uri))

    39     

    40             results = endpoint.query().convert()['results']['bindings']

    41             

    42             if len(results) == 1: # We don't want places located in multiple countries

    43                 

    44                 resourceuri = results[0]['resource']['value']

    45                 

    46                 if re.match(resourceprefix, resourceuri):

    47                     countrytxt = re.findall('([^/]+$)', resourceuri)[0]

    48                     

    49                     country, created = Country.objects.get_or_create(dbpedia_uri=resourceuri)

    50                     GeoInclusion.objects.get_or_create(tag=tag, country=country)

    51                     

    52                 if resourceuri == identityuri:

    53                     countrytxt = '<is a country>'

    54                     

    55                     country, created = Country.objects.get_or_create(dbpedia_uri=tag.dbpedia_uri)

    56                     GeoInclusion.objects.get_or_create(tag=tag, country=country)

    57                 

    58             else:

    59                 countrytxt = '<unknown>'

    60             

    61             writer = show_progress(i+1, total, '%s => %s'%(tag.label, countrytxt), 50, writer)