def test_with_distinct_counts_raises_when_queryset_includes_unsupported_options(
self):
"""
Verify that an error is raised if the original queryset includes options that are not supported by our
custom Search class.
"""
dc_queryset = DistinctCountsSearchQuerySet.from_queryset(
DSLFacetedSearch())
with pytest.raises(RuntimeError) as err:
facet_field = 'start'
agg_filter = ESDSLQ('match_all')
agg = DateHistogramFacet(field=facet_field, interval='month')
dc_queryset.aggs.bucket('_filter_' + facet_field,
'filter',
filter=agg_filter).bucket(
facet_field, agg.get_aggregation())
dc_queryset.with_distinct_counts('aggregation_key')
assert str(
err.value
) == 'DistinctCountsSearchQuerySet does not support date facets.'
dc_queryset = DistinctCountsSearchQuerySet.from_queryset(
DSLFacetedSearch())
with pytest.raises(RuntimeError) as err:
facet_field = 'pacing_type'
agg_filter = ESDSLQ('match_all')
agg = TermsFacet(field=facet_field, order='term')
dc_queryset.aggs.bucket('_filter_' + facet_field,
'filter',
filter=agg_filter).bucket(
facet_field, agg.get_aggregation())
dc_queryset.with_distinct_counts('aggregation_key')
assert 'DistinctCountsSearchQuerySet only supports a limited set of field facet options.' in str(
err.value)
def test_distinct_count_returns_cached_distinct_count(self):
""" Verify that distinct_count returns the cached distinct_result_count when present."""
queryset = DSLFacetedSearch()
dc_queryset = DistinctCountsSearchQuerySet.from_queryset(
queryset).with_distinct_counts('aggregation_key')
dc_queryset._distinct_result_count = 123 # pylint: disable=protected-access
assert dc_queryset.distinct_count() == 123
def test_facet_counts_caches_results(self):
""" Verify that facet_counts cache results when it is forced to run the query."""
course = CourseFactory()
runs = [
CourseRunFactory(title='foo',
pacing_type='self_paced',
hidden=True,
course=course),
CourseRunFactory(title='foo',
pacing_type='self_paced',
hidden=True,
course=course),
CourseRunFactory(title='foo',
pacing_type='instructor_paced',
hidden=False,
course=course),
]
queryset = DSLFacetedSearch(
index=CourseRunDocument._index._name).filter('term', title='foo')
facet_field = 'pacing_type'
agg_filter = ESDSLQ('match_all')
agg = TermsFacet(field=facet_field)
queryset.aggs.bucket('_filter_' + facet_field,
'filter',
filter=agg_filter).bucket(facet_field,
agg.get_aggregation())
queryset.aggs.bucket('_query_{0}'.format('hidden'),
'filter',
filter=ESDSLQ('bool',
filter=ESDSLQ('term', hidden=True)))
dc_queryset = DistinctCountsSearchQuerySet.from_queryset(
queryset).with_distinct_counts('aggregation_key')
# This should force the query to execute, and the results to be cached
facet_counts = dc_queryset.facet_counts()
with mock.patch.object(DistinctCountsElasticsearchQueryWrapper,
'search') as mock_search:
# Calling facet_counts again shouldn't result in an additional query
cached_facet_counts = dc_queryset.facet_counts()
assert not mock_search.called
assert facet_counts == cached_facet_counts
# Calling count shouldn't result in another query, as we should have already cached it with the
# first request.
count = dc_queryset.count()
assert not mock_search.called
assert count == len(runs)
# Fetching the results shouldn't result in another query, as we should have already cached them
# with the initial request.
results = dc_queryset.execute()
assert not mock_search.called
expected = {run.key for run in runs}
actual = {run.key for run in results}
assert expected == actual
def test_with_distinct_counts(self):
"""
Verify that the query object is converted to a DistinctCountsSearchQuerySet and the aggregation_key is
configured properly.
"""
queryset = DSLFacetedSearch()
dc_queryset = DistinctCountsSearchQuerySet.from_queryset(
queryset).with_distinct_counts('aggregation_key')
assert isinstance(dc_queryset, DistinctCountsSearchQuerySet)
assert dc_queryset.aggregation_key == 'aggregation_key'
def test_distinct_count_raises_when_not_properly_configured(self):
"""
Verify that distinct_count raises when called without configuring the DSLFacetedSearch to compute distinct
counts.
"""
queryset = DSLFacetedSearch()
dc_queryset = DistinctCountsSearchQuerySet.from_queryset(queryset)
with pytest.raises(AttributeError) as err:
dc_queryset.distinct_count()
assert "object has no attribute 'distinct_count'" in str(err.value)
def test_facet_counts_includes_distinct_counts(self):
""" Verify that facet_counts include distinct counts. """
course = CourseFactory()
CourseRunFactory(title='foo',
pacing_type='self_paced',
hidden=True,
course=course)
CourseRunFactory(title='foo',
pacing_type='self_paced',
hidden=True,
course=course)
CourseRunFactory(title='foo',
pacing_type='instructor_paced',
hidden=False,
course=course)
# Make sure to add both a field facet and a query facet so that we can be sure that both work.
queryset = DSLFacetedSearch(
index=CourseRunDocument._index._name).filter('term', title='foo')
facet_field = 'pacing_type'
agg_filter = ESDSLQ('match_all')
agg = TermsFacet(field=facet_field)
queryset.aggs.bucket('_filter_' + facet_field,
'filter',
filter=agg_filter).bucket(facet_field,
agg.get_aggregation())
queryset.aggs.bucket('_query_{0}'.format('hidden'),
'filter',
filter=ESDSLQ('bool',
filter=ESDSLQ('term', hidden=True)))
dc_queryset = DistinctCountsSearchQuerySet.from_queryset(
queryset).with_distinct_counts('aggregation_key')
facet_counts = dc_queryset.facet_counts()
# Field facets are expected to be formatted as a list of three-tuples (field_value, count, distinct_count)
for val, count, distinct_count in facet_counts['fields'][
'pacing_type']:
assert val in {'self_paced', 'instructor_paced'}
if val == 'self_paced':
assert count == 2
assert distinct_count == 1
elif val == 'instructor_paced':
assert count == 1
assert distinct_count == 1
# Query facets are expected to be formatted as a dictionary mapping facet_names to two-tuples (count,
# distinct_count)
hidden_count, hidden_distinct_count = facet_counts['queries']['hidden']
assert hidden_count == 2
assert hidden_distinct_count == 1
def test_from_queryset(self):
""" Verify that a DistinctCountsSearchQuerySet can be built from an existing SearchQuerySet."""
course_1 = CourseFactory()
CourseRunFactory(title='foo', course=course_1)
CourseRunFactory(title='foo', course=course_1)
course_2 = CourseFactory()
CourseRunFactory(title='foo', course=course_2)
CourseRunFactory(title='bar', course=course_2)
queryset = DSLFacetedSearch(
index=CourseRunDocument._index._name).filter('term', title='foo')
dc_queryset = DistinctCountsSearchQuerySet.from_queryset(queryset)
expected = sorted([run.key for run in queryset])
actual = sorted([run.key for run in dc_queryset])
assert expected == actual
def test_distinct_count_runs_query_when_cache_is_empty(self):
""" Verify that distinct_count runs the query, caches, and returns the distinct_count when cache is empty."""
course_1 = CourseFactory()
CourseRunFactory(title='foo', course=course_1)
CourseRunFactory(title='foo', course=course_1)
course_2 = CourseFactory()
CourseRunFactory(title='foo', course=course_2)
CourseRunFactory(title='bar', course=course_2)
queryset = DSLFacetedSearch(
index=CourseRunDocument._index._name).filter('term', title='foo')
dc_queryset = DistinctCountsSearchQuerySet.from_queryset(
queryset).with_distinct_counts('aggregation_key')
assert dc_queryset._distinct_result_count is None # pylint: disable=protected-access
assert dc_queryset.distinct_count() == 2
assert dc_queryset._distinct_result_count == 2 # pylint: disable=protected-access
def get_queryset(self, *args, **kwargs):
""" Return the base Queryset to use to build up the search query."""
queryset = super().get_queryset(*args, **kwargs)
return DistinctCountsSearchQuerySet.from_queryset(
queryset).with_distinct_counts('aggregation_key')
