src/p4l/management/commands/get_language_label.py
changeset 10 c4e7d66b7dc2
child 12 57efd01f1715
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/src/p4l/management/commands/get_language_label.py	Fri Aug 30 13:08:29 2013 +0200
@@ -0,0 +1,73 @@
+# -*- coding: utf-8 -*-
+
+from django.conf import settings
+from django.core.management import BaseCommand
+from p4l.models import Language
+from p4l.utils import show_progress
+import logging
+import requests
+
+logger = logging.getLogger(__name__)
+
+class Command(BaseCommand):
+    
+    query = """
+PREFIX skos:<http://www.w3.org/2004/02/skos/core#>
+PREFIX rdf:<http://www.w3.org/1999/02/22-rdf-syntax-ns#>
+PREFIX owl:<http://www.w3.org/2002/07/owl#>
+PREFIX rdfs:<http://www.w3.org/2000/01/rdf-schema#>
+SELECT ?uri ?label 
+WHERE {
+    ?uri skos:inScheme <http://www.iiep.unesco.org/plan4learning/scheme/Languages> .
+    ?uri skos:prefLabel|skos:label ?label .
+    FILTER (?uri = $root)
+}
+"""
+    
+    def fill_label(self):
+        # Loads Subjects label from sparkl query
+        langs = Language.objects.filter(label=None)
+        total_langs = len(langs)
+        writer = None
+        i = 0
+        found = 0
+        for l in langs:
+            i += 1
+            logger.debug("1")
+            logger.debug(l)
+            logger.debug("2")
+            logger.debug(l.language)
+            res = requests.get(
+                settings.SPARQL_QUERY_ENDPOINT,
+                params={'query':self.query, 'timeout':10, '$root' : "<"+l.language+">"},
+                headers={'accept':'application/sparql-results+json'},
+            )
+            if not res.ok:
+                continue
+            elif res.text:
+                json_res = res.json()
+                if 'results' in json_res and 'bindings' in json_res['results'] and len(json_res['results']['bindings'])>0:
+                    # json_res['results']['bindings'] has several languages. If we find french, we save the french label.
+                    # If not, we save the first one.
+                    tmp_dict = {}
+                    first_label = None
+                    # We create a temporary dict with the lang code and the label
+                    for b in json_res['results']['bindings']:
+                        if 'label' in b and 'value' in b['label'] and 'xml:lang' in b['label']:
+                            tmp_dict[b['label']['xml:lang']] = b['label']['value']
+                            if not first_label:
+                                first_label = b['label']['value']
+                    if 'fr' in tmp_dict or first_label:
+                        if 'fr' in tmp_dict:
+                            l.label = tmp_dict['fr']
+                        else:
+                            l.label = first_label
+                        lab = l.label
+                        l.save()
+                        found += 1
+                        writer = show_progress(i, total_langs, lab, 50, writer=writer)
+        print("Processing Subjects Sparql Done. %d found on %d" % (found, total_langs))
+
+    def handle(self, *args, **options):
+        self.fill_label()
+