def test__in_mem_dm__get_hit_rows_for_queries():
query_rome = 'rome'
query_caesar = 'caesar'
record_brutus = 'brutus'
record_caesar = 'caesar'
hits = {
query_rome: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
record_brutus: Hit(2.0, datetime(1999, 1, 1)),
},
query_caesar: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
}
}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(
return_value=hits.iteritems())
sut = InMemoryQueryDataModel(fake_model)
ret_hits = list(sut.get_hit_rows_for_queries([query_rome]))
assert len(ret_hits) == 1
query, q_hits = ret_hits[0]
assert query == query_rome
for record, hit in q_hits.iteritems():
assert hits[query_rome][record] == hit
def test__ThresholdQueryNeighbourhood():
query_string = 'fat cat'
similarities = {
'fat rat': 0.5,
'fat fat': 0.49,
}
fake_model = AbstractQueryDataModel()
fake_sim = AbstractQuerySimilarity()
def query_sim(query1, query2):
return similarities[query2]
def get_queries():
return similarities.keys()
fake_sim.get_similarity = mock.Mock(side_effect=query_sim)
fake_model.get_queries = mock.Mock(side_effect=get_queries)
sut = ThresholdQueryNeighbourhood(data_model=fake_model,
query_sim=fake_sim,
sim_threshold=0.5)
nbours = list(sut.get_neighbours(query_string))
assert len(nbours) == 1
assert nbours[0] == 'fat rat'
def test__in_mem_dm__get_hit_rows_for_queries():
query_rome = 'rome'
query_caesar = 'caesar'
record_brutus = 'brutus'
record_caesar = 'caesar'
hits = {
query_rome: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
record_brutus: Hit(2.0, datetime(1999, 1, 1)),
},
query_caesar: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
}
}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(return_value=hits.iteritems())
sut = InMemoryQueryDataModel(fake_model)
ret_hits = list(sut.get_hit_rows_for_queries([query_rome]))
assert len(ret_hits) == 1
query, q_hits = ret_hits[0]
assert query == query_rome
for record, hit in q_hits.iteritems():
assert hits[query_rome][record] == hit
def test__in_mem_dm__get_hit_rows_for_queries__unknown_query():
query_rome = 'rome'
query_caesar = 'caesar'
query_unknown = 'unknown'
record_brutus = 'brutus'
record_caesar = 'caesar'
hits = {
query_rome: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
record_brutus: Hit(2.0, datetime(1999, 1, 1)),
},
query_caesar: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
}
}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(return_value=hits.iteritems())
sut = InMemoryQueryDataModel(fake_model)
ret_hits = list(sut.get_hit_rows_for_queries([query_unknown]))
assert len(ret_hits) == 0
def test__in_mem_dm__get_hit_rows_for_queries__unknown_query():
query_rome = 'rome'
query_caesar = 'caesar'
query_unknown = 'unknown'
record_brutus = 'brutus'
record_caesar = 'caesar'
hits = {
query_rome: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
record_brutus: Hit(2.0, datetime(1999, 1, 1)),
},
query_caesar: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
}
}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(
return_value=hits.iteritems())
sut = InMemoryQueryDataModel(fake_model)
ret_hits = list(sut.get_hit_rows_for_queries([query_unknown]))
assert len(ret_hits) == 0
def test__ThresholdQueryNeighbourhood():
query_string = 'fat cat'
similarities = {
'fat rat': 0.5,
'fat fat': 0.49,
}
fake_model = AbstractQueryDataModel()
fake_sim = AbstractQuerySimilarity()
def query_sim(query1, query2):
return similarities[query2]
def get_queries():
return similarities.keys()
fake_sim.get_similarity = mock.Mock(side_effect=query_sim)
fake_model.get_queries = mock.Mock(side_effect=get_queries)
sut = ThresholdQueryNeighbourhood(
data_model=fake_model,
query_sim=fake_sim,
sim_threshold=0.5)
nbours = list(sut.get_neighbours(query_string))
assert len(nbours) == 1
assert nbours[0] == 'fat rat'
def test__in_mem_dm__get_hit_rows():
query_rome = 'rome'
query_caesar = 'caesar'
record_brutus = 'brutus'
record_caesar = 'caesar'
hits = {
query_rome: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
record_brutus: Hit(2.0, datetime(1999, 1, 1)),
},
query_caesar: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
}
}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(
return_value=hits.iteritems())
sut = InMemoryQueryDataModel(fake_model)
for query, q_hits in sut.get_hit_rows():
for record, hit in q_hits.iteritems():
assert hits[query][record] == hit
def test__in_mem_dm__refresh__underlying_data_model_is_refreshed():
hits = {}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(return_value=hits.iteritems())
fake_model.refresh = mock.Mock()
sut = InMemoryQueryDataModel(fake_model)
refreshed_components = set()
sut.refresh(refreshed_components)
fake_model.refresh.assert_called_once_with(refreshed_components)
assert fake_model in refreshed_components
assert sut in refreshed_components
def test__in_mem_dm__refresh__underlying_data_model_is_refreshed():
hits = {
}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(
return_value=hits.iteritems())
fake_model.refresh = mock.Mock()
sut = InMemoryQueryDataModel(fake_model)
refreshed_components = set()
sut.refresh(refreshed_components)
fake_model.refresh.assert_called_once_with(refreshed_components)
assert fake_model in refreshed_components
assert sut in refreshed_components
def test__thres_nhood__refresh__data_model_and_similarity_refreshed():
fake_model = AbstractQueryDataModel()
fake_model.refresh = mock.Mock()
fake_sim = AbstractQuerySimilarity()
fake_sim.refresh = mock.Mock()
sut = ThresholdQueryNeighbourhood(data_model=fake_model,
query_sim=fake_sim,
sim_threshold=0.5)
refreshed_components = set()
sut.refresh(refreshed_components)
assert fake_model in refreshed_components
assert fake_sim in refreshed_components
assert sut in refreshed_components
assert fake_model.refresh.call_count == 1
assert fake_sim.refresh.call_count == 1
def test__thres_nhood__refresh__data_model_and_similarity_refreshed():
fake_model = AbstractQueryDataModel()
fake_model.refresh = mock.Mock()
fake_sim = AbstractQuerySimilarity()
fake_sim.refresh = mock.Mock()
sut = ThresholdQueryNeighbourhood(
data_model=fake_model,
query_sim=fake_sim,
sim_threshold=0.5)
refreshed_components = set()
sut.refresh(refreshed_components)
assert fake_model in refreshed_components
assert fake_sim in refreshed_components
assert sut in refreshed_components
assert fake_model.refresh.call_count == 1
assert fake_sim.refresh.call_count == 1
def test__in_mem_dm__refresh__data_is_reloaded():
query_rome = 'rome'
record_caesar = 'caesar'
hits = {}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(return_value=hits.iteritems())
fake_model.refresh = mock.Mock()
sut = InMemoryQueryDataModel(fake_model)
assert {k: v for k, v in sut.get_hit_rows()} == {}
hits[query_rome] = {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
}
sut.refresh(set())
for record_id, rec_prefs in sut.get_hit_rows():
for session_id, pref in rec_prefs.iteritems():
assert hits[record_id][session_id] == pref
def test__in_mem_dm__get_queries():
query_rome = 'rome'
query_caesar = 'caesar'
record_brutus = 'brutus'
record_caesar = 'caesar'
hits = {
query_rome: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
record_brutus: Hit(2.0, datetime(1999, 1, 1)),
},
query_caesar: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
}
}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(return_value=hits.iteritems())
sut = InMemoryQueryDataModel(fake_model)
ret_queries = list(sut.get_queries())
assert sorted(hits.keys()) == sorted(ret_queries)
def test__in_mem_dm__refresh__data_is_reloaded():
query_rome = 'rome'
record_caesar = 'caesar'
hits = {}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(
return_value=hits.iteritems())
fake_model.refresh = mock.Mock()
sut = InMemoryQueryDataModel(fake_model)
assert {k: v for k, v in sut.get_hit_rows()} == {}
hits[query_rome] = {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
}
sut.refresh(set())
for record_id, rec_prefs in sut.get_hit_rows():
for session_id, pref in rec_prefs.iteritems():
assert hits[record_id][session_id] == pref
def test__q_recommender__refresh__underlying_components_are_refreshed():
fake_model = AbstractQueryDataModel()
fake_model.refresh = mock.Mock()
fake_sim = AbstractQuerySimilarity()
fake_sim.refresh = mock.Mock()
fake_nhood = AbstractQueryNeighbourhood()
fake_nhood.refresh = mock.Mock()
sut = QueryBasedRecommender(
data_model=fake_model,
query_sim=fake_sim,
query_nhood=fake_nhood,
scorer=WeightedSumScorer(lambda r, hit_row: 1.0))
refreshed_components = set()
sut.refresh(refreshed_components)
assert fake_model in refreshed_components
assert fake_sim in refreshed_components
assert fake_nhood in refreshed_components
assert sut in refreshed_components
assert fake_model.refresh.call_count == 1
assert fake_sim.refresh.call_count == 1
assert fake_nhood.refresh.call_count == 1
def test__in_mem_dm__get_queries():
query_rome = 'rome'
query_caesar = 'caesar'
record_brutus = 'brutus'
record_caesar = 'caesar'
hits = {
query_rome: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
record_brutus: Hit(2.0, datetime(1999, 1, 1)),
},
query_caesar: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
}
}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(
return_value=hits.iteritems())
sut = InMemoryQueryDataModel(fake_model)
ret_queries = list(sut.get_queries())
assert sorted(hits.keys()) == sorted(ret_queries)
def test__q_recommender__refresh__underlying_components_are_refreshed():
fake_model = AbstractQueryDataModel()
fake_model.refresh = mock.Mock()
fake_sim = AbstractQuerySimilarity()
fake_sim.refresh = mock.Mock()
fake_nhood = AbstractQueryNeighbourhood()
fake_nhood.refresh = mock.Mock()
sut = QueryBasedRecommender(
data_model=fake_model,
query_sim=fake_sim,
query_nhood=fake_nhood,
scorer=WeightedSumScorer(lambda r, hit_row: 1.0))
refreshed_components = set()
sut.refresh(refreshed_components)
assert fake_model in refreshed_components
assert fake_sim in refreshed_components
assert fake_nhood in refreshed_components
assert sut in refreshed_components
assert fake_model.refresh.call_count == 1
assert fake_sim.refresh.call_count == 1
assert fake_nhood.refresh.call_count == 1
def test__in_mem_dm__get_hit_rows():
query_rome = 'rome'
query_caesar = 'caesar'
record_brutus = 'brutus'
record_caesar = 'caesar'
hits = {
query_rome: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
record_brutus: Hit(2.0, datetime(1999, 1, 1)),
},
query_caesar: {
record_caesar: Hit(1.0, datetime(1999, 1, 1)),
}
}
fake_model = AbstractQueryDataModel()
fake_model.get_hit_rows = mock.Mock(return_value=hits.iteritems())
sut = InMemoryQueryDataModel(fake_model)
for query, q_hits in sut.get_hit_rows():
for record, hit in q_hits.iteritems():
assert hits[query][record] == hit