def get(self, request, *args, **kwargs):
limit = request.GET.get('limit', 5)
es_query = {
'apps': {
'completion': {
'field': 'name_suggest',
'size': limit
},
'text': request.GET.get('q', '').strip()
}
}
results = WebappIndexer.get_es().suggest(
body=es_query, index=WebappIndexer.get_index())
if 'apps' in results:
data = results['apps'][0]['options']
else:
data = []
serializer = self.get_serializer(data)
# This returns a JSON list. Usually this is a bad idea for security
# reasons, but we don't include any user-specific data, it's fully
# anonymous, so we're fine.
return HttpResponse(json.dumps(serializer.data),
content_type='application/x-rocketbar+json')
def tearDown(self):
# Cleanup to remove these from the index.
self.app1.delete()
self.app2.delete()
unindex_webapps([self.app1.id, self.app2.id])
# Required to purge the suggestions data structure. In Lucene, a
# document is not deleted from a segment, just marked as deleted.
WebappIndexer.get_es().indices.optimize(index=WebappIndexer.get_index(), only_expunge_deletes=True)
def tearDown(self):
# Cleanup to remove these from the index.
self.app1.delete()
self.app2.delete()
unindex_webapps([self.app1.id, self.app2.id])
# Required to purge the suggestions data structure. In Lucene, a
# document is not deleted from a segment, just marked as deleted.
WebappIndexer.get_es().indices.optimize(
index=WebappIndexer.get_index(), only_expunge_deletes=True)
def _get_indices(self):
# Check if we are filtering by a doc_type (e.g., apps, sites).
# Default to all content types.
doc_type = self.request.GET.get('doc_type', 'all')
app_index = WebappIndexer.get_index()
site_index = WebsiteIndexer.get_index()
if doc_type == 'webapp':
return [app_index]
elif doc_type == 'website':
return [site_index]
return [app_index, site_index]
def _get_indices(self):
# Check if we are filtering by a doc_type (e.g., apps, sites).
# Default to all content types.
doc_type = self.request.GET.get('doc_type', 'all')
app_index = WebappIndexer.get_index()
site_index = WebsiteIndexer.get_index()
if doc_type == 'webapp':
return [app_index]
elif doc_type == 'website':
return [site_index]
return [app_index, site_index]
def mget_apps(self, app_ids):
"""
Takes a list of app_ids. Does an ES mget.
Returns an app_map for serializer context.
"""
app_map = {}
es = WebappIndexer.get_es()
apps = es.mget(body={'ids': app_ids}, index=WebappIndexer.get_index(),
doc_type=WebappIndexer.get_mapping_type_name())
for app in apps['docs']:
# Store the apps to attach to feed elements later.
app = app['_source']
app_map[app['id']] = app
return app_map
def index_webapps(ids, **kw):
"""TODO: use search/indexers.py:index."""
task_log.info('Indexing apps %s-%s. [%s]' % (ids[0], ids[-1], len(ids)))
index = kw.pop('index', WebappIndexer.get_index())
# Note: If reindexing is currently occurring, `get_indices` will return
# more than one index.
indices = Reindexing.get_indices(index)
es = WebappIndexer.get_es(urls=settings.ES_URLS)
qs = Webapp.indexing_transformer(Webapp.with_deleted.no_cache().filter(
id__in=ids))
for obj in qs:
doc = WebappIndexer.extract_document(obj.id, obj)
for idx in indices:
WebappIndexer.index(doc, id_=obj.id, es=es, index=idx)
def unindex_webapps(ids, **kw):
if not ids:
return
task_log.info('Un-indexing apps %s-%s. [%s]' % (ids[0], ids[-1], len(ids)))
index = kw.pop('index', WebappIndexer.get_index())
# Note: If reindexing is currently occurring, `get_indices` will return
# more than one index.
indices = Reindexing.get_indices(index)
es = WebappIndexer.get_es(urls=settings.ES_URLS)
for id_ in ids:
for idx in indices:
try:
WebappIndexer.unindex(id_=id_, es=es, index=idx)
except ElasticHttpNotFoundError:
# Ignore if it's not there.
task_log.info(
u'[Webapp:%s] Unindexing app but not found in index' % id_)
def get(self, request, *args, **kwargs):
limit = request.GET.get('limit', 5)
es_query = {
'apps': {
'completion': {'field': 'name_suggest', 'size': limit},
'text': request.GET.get('q', '').strip()
}
}
results = S(WebappIndexer).get_es().send_request(
'GET', [WebappIndexer.get_index(), '_suggest'], body=es_query)
if 'apps' in results:
data = results['apps'][0]['options']
else:
data = []
serializer = self.get_serializer(data)
# This returns a JSON list. Usually this is a bad idea for security
# reasons, but we don't include any user-specific data, it's fully
# anonymous, so we're fine.
return HttpResponse(json.dumps(serializer.data),
content_type='application/x-rocketbar+json')
def handle(self, *args, **kwargs):
index = WebappIndexer.get_index()
doctype = WebappIndexer.get_mapping_type_name()
es = WebappIndexer.get_es()
app_ids = Webapp.objects.values_list('id', flat=True)
missing_ids = []
for app_id in app_ids:
try:
es.get(index, app_id, doctype, fields='id')
except elasticsearch.NotFoundError:
# App doesn't exist in our index, add it to `missing_ids`.
missing_ids.append(app_id)
if missing_ids:
sys.stdout.write('Adding %s doc(s) to the index.'
% len(missing_ids))
WebappIndexer().run_indexing(missing_ids, es)
else:
sys.stdout.write('No docs missing from index.')
def handle(self, *args, **kwargs):
index = WebappIndexer.get_index()
doctype = WebappIndexer.get_mapping_type_name()
es = WebappIndexer.get_es()
app_ids = Webapp.objects.values_list('id', flat=True)
missing_ids = []
for app_id in app_ids:
try:
es.get(index, app_id, doctype, fields='id')
except elasticsearch.NotFoundError:
# App doesn't exist in our index, add it to `missing_ids`.
missing_ids.append(app_id)
if missing_ids:
sys.stdout.write('Adding %s doc(s) to the index.' %
len(missing_ids))
WebappIndexer().run_indexing(missing_ids, es)
else:
sys.stdout.write('No docs missing from index.')
def test_index(self):
with self.settings(ES_INDEXES={'webapp': 'apps'}):
eq_(WebappIndexer.get_index(), 'apps')
def test_index(self):
with self.settings(ES_INDEXES={'webapp': 'apps'}):
eq_(WebappIndexer.get_index(), 'apps')
def test_index(self):
with self.settings(ES_INDEXES={"webapp": "apps"}):
eq_(WebappIndexer.get_index(), "apps")
def get_queryset(self):
return Search(
using=BaseIndexer.get_es(),
index=[WebappIndexer.get_index(), WebsiteIndexer.get_index()],
doc_type=[WebappIndexer.get_mapping_type_name(),
WebsiteIndexer.get_mapping_type_name()])
