Skip to content

Commit 6ff22c9

Browse files
joeinjoein
committed
deprecate: replace recommend with query points in test sparse recommend
1 parent d46d0d0 commit 6ff22c9

File tree

1 file changed

+134
-105
lines changed

1 file changed

+134
-105
lines changed

‎tests/congruence_tests/test_sparse_recommend.py‎

Lines changed: 134 additions & 105 deletions
Original file line numberDiff line numberDiff line change
@@ -30,226 +30,251 @@ def __init__(self):
3030

3131
@classmethod
3232
def simple_recommend_image(cls, client: QdrantBase) -> list[models.ScoredPoint]:
33-
return client.recommend(
33+
return client.query_points(
3434
collection_name=COLLECTION_NAME,
35-
positive=[10],
36-
negative=[],
35+
query=models.RecommendQuery(
36+
recommend=models.RecommendInput(positive=[10], negative=[])
37+
),
3738
with_payload=True,
3839
limit=10,
3940
using="sparse-image",
40-
)
41+
).points
4142

4243
@classmethod
4344
def many_recommend(cls, client: QdrantBase) -> list[models.ScoredPoint]:
44-
return client.recommend(
45+
return client.query_points(
4546
collection_name=COLLECTION_NAME,
46-
positive=[10, 19],
47+
query=models.RecommendQuery(recommend=models.RecommendInput(positive=[10, 19])),
4748
with_payload=True,
4849
limit=10,
4950
using="sparse-image",
50-
)
51+
).points
5152

5253
@classmethod
5354
def simple_recommend_negative(cls, client: QdrantBase) -> list[models.ScoredPoint]:
54-
return client.recommend(
55+
return client.query_points(
5556
collection_name=COLLECTION_NAME,
56-
positive=[10],
57-
negative=[15, 7],
57+
query=models.RecommendQuery(
58+
recommend=models.RecommendInput(positive=[10], negative=[15, 7])
59+
),
5860
with_payload=True,
5961
limit=10,
6062
using="sparse-image",
61-
)
63+
).points
6264

6365
@classmethod
6466
def recommend_from_another_collection(cls, client: QdrantBase) -> list[models.ScoredPoint]:
65-
return client.recommend(
67+
return client.query_points(
6668
collection_name=COLLECTION_NAME,
67-
positive=[10],
68-
negative=[15, 7],
69+
query=models.RecommendQuery(
70+
recommend=models.RecommendInput(positive=[10], negative=[15, 7])
71+
),
6972
with_payload=True,
7073
limit=10,
7174
using="sparse-image",
7275
lookup_from=models.LookupLocation(
7376
collection=secondary_collection_name,
7477
vector="sparse-image",
7578
),
76-
)
79+
).points
7780

7881
@classmethod
7982
def filter_recommend_text(
8083
cls, client: QdrantBase, query_filter: models.Filter
8184
) -> list[models.ScoredPoint]:
82-
return client.recommend(
85+
return client.query_points(
8386
collection_name=COLLECTION_NAME,
84-
positive=[10],
87+
query=models.RecommendQuery(recommend=models.RecommendInput(positive=[10])),
8588
query_filter=query_filter,
8689
with_payload=True,
8790
limit=10,
8891
using="sparse-text",
89-
)
92+
).points
9093

9194
@classmethod
9295
def best_score_recommend(cls, client: QdrantBase) -> list[models.ScoredPoint]:
93-
return client.recommend(
96+
return client.query_points(
9497
collection_name=COLLECTION_NAME,
95-
positive=[
96-
10,
97-
20,
98-
],
99-
negative=[],
98+
query=models.RecommendQuery(
99+
recommend=models.RecommendInput(
100+
positive=[10, 20], negative=[], strategy=models.RecommendStrategy.BEST_SCORE
101+
)
102+
),
100103
with_payload=True,
101104
limit=10,
102105
using="sparse-image",
103-
strategy=models.RecommendStrategy.BEST_SCORE,
104-
)
106+
).points
105107

106108
@classmethod
107109
def best_score_recommend_euclid(cls, client: QdrantBase) -> list[models.ScoredPoint]:
108-
return client.recommend(
110+
return client.query_points(
109111
collection_name=COLLECTION_NAME,
110-
positive=[
111-
10,
112-
20,
113-
],
114-
negative=[11, 21],
112+
query=models.RecommendQuery(
113+
recommend=models.RecommendInput(
114+
positive=[10, 20],
115+
negative=[11, 21],
116+
strategy=models.RecommendStrategy.BEST_SCORE,
117+
)
118+
),
115119
with_payload=True,
116120
limit=10,
117121
using="sparse-code",
118-
strategy=models.RecommendStrategy.BEST_SCORE,
119-
)
122+
).points
120123

121124
@classmethod
122125
def only_negatives_best_score_recommend(cls, client: QdrantBase) -> list[models.ScoredPoint]:
123-
return client.recommend(
126+
return client.query_points(
124127
collection_name=COLLECTION_NAME,
125-
positive=None,
126-
negative=[10, 12],
128+
query=models.RecommendQuery(
129+
recommend=models.RecommendInput(
130+
positive=None, negative=[10, 12], strategy=models.RecommendStrategy.BEST_SCORE
131+
)
132+
),
127133
with_payload=True,
128134
limit=10,
129135
using="sparse-image",
130-
strategy=models.RecommendStrategy.BEST_SCORE,
131-
)
136+
).points
132137

133138
@classmethod
134139
def only_negatives_best_score_recommend_euclid(
135140
cls, client: QdrantBase
136141
) -> list[models.ScoredPoint]:
137-
return client.recommend(
142+
return client.query_points(
138143
collection_name=COLLECTION_NAME,
139-
positive=None,
140-
negative=[10, 12],
144+
query=models.RecommendQuery(
145+
recommend=models.RecommendInput(
146+
positive=None, negative=[10, 12], strategy=models.RecommendStrategy.BEST_SCORE
147+
)
148+
),
141149
with_payload=True,
142150
limit=10,
143151
using="sparse-code",
144-
strategy=models.RecommendStrategy.BEST_SCORE,
145-
)
152+
).points
146153

147154
@classmethod
148155
def sum_scores_recommend(cls, client: QdrantBase) -> list[models.ScoredPoint]:
149-
return client.recommend(
156+
return client.query_points(
150157
collection_name=COLLECTION_NAME,
151-
positive=[
152-
10,
153-
20,
154-
],
155-
negative=[],
158+
query=models.RecommendQuery(
159+
recommend=models.RecommendInput(
160+
positive=[10, 20], negative=[], strategy=models.RecommendStrategy.SUM_SCORES
161+
)
162+
),
156163
with_payload=True,
157164
limit=10,
158165
using="sparse-image",
159-
strategy=models.RecommendStrategy.SUM_SCORES,
160-
)
166+
).points
161167

162168
@classmethod
163169
def sum_scores_recommend_euclid(cls, client: QdrantBase) -> list[models.ScoredPoint]:
164-
return client.recommend(
170+
return client.query_points(
165171
collection_name=COLLECTION_NAME,
166-
positive=[
167-
10,
168-
20,
169-
],
170-
negative=[11, 21],
172+
query=models.RecommendQuery(
173+
recommend=models.RecommendInput(
174+
positive=[10, 20],
175+
negative=[11, 21],
176+
strategy=models.RecommendStrategy.SUM_SCORES,
177+
)
178+
),
171179
with_payload=True,
172180
limit=10,
173181
using="sparse-code",
174-
strategy=models.RecommendStrategy.SUM_SCORES,
175-
)
182+
).points
176183

177184
@classmethod
178185
def only_negatives_sum_scores_recommend(cls, client: QdrantBase) -> list[models.ScoredPoint]:
179-
return client.recommend(
186+
return client.query_points(
180187
collection_name=COLLECTION_NAME,
181-
positive=None,
182-
negative=[10, 12],
188+
query=models.RecommendQuery(
189+
recommend=models.RecommendInput(
190+
positive=None, negative=[10, 12], strategy=models.RecommendStrategy.SUM_SCORES
191+
)
192+
),
183193
with_payload=True,
184194
limit=10,
185195
using="sparse-image",
186-
strategy=models.RecommendStrategy.SUM_SCORES,
187-
)
196+
).points
188197

189198
@classmethod
190199
def only_negatives_sum_scores_recommend_euclid(
191200
cls, client: QdrantBase
192201
) -> list[models.ScoredPoint]:
193-
return client.recommend(
202+
return client.query_points(
194203
collection_name=COLLECTION_NAME,
195-
positive=None,
196-
negative=[10, 12],
204+
query=models.RecommendQuery(
205+
recommend=models.RecommendInput(
206+
positive=None, negative=[10, 12], strategy=models.RecommendStrategy.SUM_SCORES
207+
)
208+
),
197209
with_payload=True,
198210
limit=10,
199211
using="sparse-code",
200-
strategy=models.RecommendStrategy.SUM_SCORES,
201-
)
212+
).points
202213

203214
@classmethod
204215
def avg_vector_recommend(cls, client: QdrantBase) -> list[models.ScoredPoint]:
205-
return client.recommend(
216+
return client.query_points(
206217
collection_name=COLLECTION_NAME,
207-
positive=[10, 13],
208-
negative=[],
218+
query=models.RecommendQuery(
219+
recommend=models.RecommendInput(
220+
positive=[10, 13],
221+
negative=[],
222+
strategy=models.RecommendStrategy.AVERAGE_VECTOR,
223+
)
224+
),
209225
with_payload=True,
210226
limit=10,
211227
using="sparse-image",
212-
strategy=models.RecommendStrategy.AVERAGE_VECTOR,
213-
)
228+
).points
214229

215230
def recommend_from_raw_vectors(self, client: QdrantBase) -> list[models.ScoredPoint]:
216-
return client.recommend(
231+
return client.query_points(
217232
collection_name=COLLECTION_NAME,
218-
positive=[self.query_image],
219-
negative=[],
233+
query=models.RecommendQuery(
234+
recommend=models.RecommendInput(positive=[self.query_image], negative=[])
235+
),
220236
with_payload=True,
221237
limit=10,
222238
using="sparse-image",
223-
)
239+
).points
224240

225241
def recommend_from_raw_vectors_and_ids(self, client: QdrantBase) -> list[models.ScoredPoint]:
226-
return client.recommend(
242+
return client.query_points(
227243
collection_name=COLLECTION_NAME,
228-
positive=[self.query_image, 10],
229-
negative=[],
244+
query=models.RecommendQuery(
245+
recommend=models.RecommendInput(positive=[self.query_image, 10], negative=[])
246+
),
230247
with_payload=True,
231248
limit=10,
232249
using="sparse-image",
233-
)
250+
).points
234251

235252
@staticmethod
236-
def recommend_batch(client: QdrantBase) -> list[list[models.ScoredPoint]]:
237-
return client.recommend_batch(
253+
def recommend_batch(client: QdrantBase) -> list[models.QueryResponse]:
254+
return client.query_batch_points(
238255
collection_name=COLLECTION_NAME,
239256
requests=[
240-
models.RecommendRequest(
241-
positive=[3],
242-
negative=[],
257+
models.QueryRequest(
258+
query=models.RecommendQuery(
259+
recommend=models.RecommendInput(
260+
positive=[3],
261+
negative=[],
262+
strategy=models.RecommendStrategy.AVERAGE_VECTOR,
263+
)
264+
),
243265
limit=1,
244266
using="sparse-image",
245-
strategy=models.RecommendStrategy.AVERAGE_VECTOR,
246267
),
247-
models.RecommendRequest(
248-
positive=[10],
249-
negative=[],
268+
models.QueryRequest(
269+
query=models.RecommendQuery(
270+
recommend=models.RecommendInput(
271+
positive=[10],
272+
negative=[],
273+
strategy=models.RecommendStrategy.BEST_SCORE,
274+
)
275+
),
250276
limit=2,
251277
using="sparse-image",
252-
strategy=models.RecommendStrategy.BEST_SCORE,
253278
lookup_from=models.LookupLocation(
254279
collection=secondary_collection_name,
255280
vector="sparse-image",
@@ -362,33 +387,37 @@ def test_query_with_nan():
362387
)
363388

364389
with pytest.raises(AssertionError):
365-
local_client.recommend(
390+
local_client.query_points(
366391
collection_name=COLLECTION_NAME,
367-
positive=[sparse_vector],
368-
negative=[],
392+
query=models.RecommendQuery(
393+
recommend=models.RecommendInput(positive=[sparse_vector], negative=[])
394+
),
369395
using=using,
370396
)
371397

372398
with pytest.raises(UnexpectedResponse):
373-
remote_client.recommend(
399+
remote_client.query_points(
374400
collection_name=COLLECTION_NAME,
375-
positive=[sparse_vector],
376-
negative=[],
401+
query=models.RecommendQuery(
402+
recommend=models.RecommendInput(positive=[sparse_vector], negative=[])
403+
),
377404
using=using,
378405
)
379406

380407
with pytest.raises(AssertionError):
381-
local_client.recommend(
408+
local_client.query_points(
382409
collection_name=COLLECTION_NAME,
383-
positive=[1],
384-
negative=[sparse_vector],
410+
query=models.RecommendQuery(
411+
recommend=models.RecommendInput(positive=[1], negative=[sparse_vector])
412+
),
385413
using=using,
386414
)
387415

388416
with pytest.raises(UnexpectedResponse):
389-
remote_client.recommend(
417+
remote_client.query_points(
390418
collection_name=COLLECTION_NAME,
391-
positive=[1],
392-
negative=[sparse_vector],
419+
query=models.RecommendQuery(
420+
recommend=models.RecommendInput(positive=[1], negative=[sparse_vector])
421+
),
393422
using=using,
394423
)

0 commit comments

Comments
 (0)