Hopiu/wagtail
1
0
Fork 0
mirror of https://github.com/Hopiu/wagtail.git synced 2026-05-24 14:53:45 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge pull request #773 from kaedroho/search/analyser-fix

Browse source 
Don't use edgengram as query analyser
This commit is contained in:
Matt Westcott 2014-10-31 10:57:20 +00:00
commit 03ead75791
2 changed files with 40 additions and 8 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
wagtail/wagtailsearch/backends/elasticsearch.py
View file

@ -62,7 +62,7 @@ class ElasticSearchMapping(object):
mapping['boost'] = field.boost
if field.partial_match:
mapping['analyzer'] = 'edgengram_analyzer'
mapping['index_analyzer'] = 'edgengram_analyzer'
mapping['include_in_all'] = True
elif isinstance(field, FilterField):
@ -80,7 +80,7 @@ class ElasticSearchMapping(object):
fields = {
'pk': dict(type='string', index='not_analyzed', store='yes', include_in_all=False),
'content_type': dict(type='string', index='not_analyzed', include_in_all=False),
'_partials': dict(type='string', analyzer='edgengram_analyzer', include_in_all=False),
'_partials': dict(type='string', index_analyzer='edgengram_analyzer', include_in_all=False),
}
fields.update(dict(

44
wagtail/wagtailsearch/tests/test_elasticsearch_backend.py
View file

@ -89,7 +89,7 @@ class TestElasticSearchBackend(BackendTests, TestCase):
# Add some test data
obj = models.SearchTest()
obj.title = "Ĥéỻø"
obj.title = "Ĥéllø"
obj.live = True
obj.save()
self.backend.add(obj)
@ -103,6 +103,38 @@ class TestElasticSearchBackend(BackendTests, TestCase):
self.assertEqual(len(results), 1)
self.assertEqual(results[0].id, obj.id)
def test_query_analyser(self):
"""
This is testing that fields that use edgengram_analyzer as their index analyser do not
have it also as their query analyser
"""
# Reset the index
self.backend.reset_index()
self.backend.add_type(models.SearchTest)
self.backend.add_type(models.SearchTestChild)
# Add some test data
obj = models.SearchTest()
obj.title = "Hello"
obj.live = True
obj.save()
self.backend.add(obj)
# Refresh the index
self.backend.refresh_index()
# Test search for "Hello"
results = self.backend.search("Hello", models.SearchTest.objects.all())
# Should find the result
self.assertEqual(len(results), 1)
# Test search for "Horse"
results = self.backend.search("Horse", models.SearchTest.objects.all())
# Even though they both start with the letter "H". This should not be considered a match
self.assertEqual(len(results), 0)
class TestElasticSearchQuery(TestCase):
def assertDictEqual(self, a, b):
@ -310,10 +342,10 @@ class TestElasticSearchMapping(TestCase):
'properties': {
'pk': {'index': 'not_analyzed', 'type': 'string', 'store': 'yes', 'include_in_all': False},
'content_type': {'index': 'not_analyzed', 'type': 'string', 'include_in_all': False},
'_partials': {'analyzer': 'edgengram_analyzer', 'include_in_all': False, 'type': 'string'},
'_partials': {'index_analyzer': 'edgengram_analyzer', 'include_in_all': False, 'type': 'string'},
'live_filter': {'index': 'not_analyzed', 'type': 'boolean', 'include_in_all': False},
'published_date_filter': {'index': 'not_analyzed', 'type': 'date', 'include_in_all': False},
'title': {'type': 'string', 'include_in_all': True, 'analyzer': 'edgengram_analyzer'},
'title': {'type': 'string', 'include_in_all': True, 'index_analyzer': 'edgengram_analyzer'},
'title_filter': {'index': 'not_analyzed', 'type': 'string', 'include_in_all': False},
'content': {'type': 'string', 'include_in_all': True},
'callable_indexed_field': {'type': 'string', 'include_in_all': True}
@ -382,15 +414,15 @@ class TestElasticSearchMappingInheritance(TestCase):
'properties': {
# New
'extra_content': {'type': 'string', 'include_in_all': True},
'subtitle': {'type': 'string', 'include_in_all': True, 'analyzer': 'edgengram_analyzer'},
'subtitle': {'type': 'string', 'include_in_all': True, 'index_analyzer': 'edgengram_analyzer'},
# Inherited
'pk': {'index': 'not_analyzed', 'type': 'string', 'store': 'yes', 'include_in_all': False},
'content_type': {'index': 'not_analyzed', 'type': 'string', 'include_in_all': False},
'_partials': {'analyzer': 'edgengram_analyzer', 'include_in_all': False, 'type': 'string'},
'_partials': {'index_analyzer': 'edgengram_analyzer', 'include_in_all': False, 'type': 'string'},
'live_filter': {'index': 'not_analyzed', 'type': 'boolean', 'include_in_all': False},
'published_date_filter': {'index': 'not_analyzed', 'type': 'date', 'include_in_all': False},
'title': {'type': 'string', 'include_in_all': True, 'analyzer': 'edgengram_analyzer'},
'title': {'type': 'string', 'include_in_all': True, 'index_analyzer': 'edgengram_analyzer'},
'title_filter': {'index': 'not_analyzed', 'type': 'string', 'include_in_all': False},
'content': {'type': 'string', 'include_in_all': True},
'callable_indexed_field': {'type': 'string', 'include_in_all': True}