deprecate python VectorQuery (#267)
* deprecate python VectorQuery * remove import
This commit is contained in:
@@ -15,7 +15,7 @@ from zvec import (
|
||||
CollectionSchema,
|
||||
Collection,
|
||||
Doc,
|
||||
VectorQuery,
|
||||
Query,
|
||||
)
|
||||
|
||||
from support_helper import *
|
||||
|
||||
@@ -833,7 +833,7 @@ class TestIndexDDL:
|
||||
)
|
||||
|
||||
query_result: list[Doc] = coll.query(
|
||||
VectorQuery(field_name=vector_name, vector=query_vector),
|
||||
Query(field_name=vector_name, vector=query_vector),
|
||||
include_vector=False,
|
||||
topk=len(docs),
|
||||
)
|
||||
|
||||
@@ -11,7 +11,7 @@ from zvec import (
|
||||
CollectionSchema,
|
||||
Collection,
|
||||
Doc,
|
||||
VectorQuery,
|
||||
Query,
|
||||
StatusCode,
|
||||
)
|
||||
from distance_helper import *
|
||||
@@ -456,7 +456,7 @@ def singledoc_and_check(
|
||||
for k, v in DEFAULT_VECTOR_FIELD_NAME.items():
|
||||
if v != {}:
|
||||
query_result = collection.query(
|
||||
VectorQuery(field_name=v, vector=insert_doc.vectors[v]),
|
||||
Query(field_name=v, vector=insert_doc.vectors[v]),
|
||||
topk=10,
|
||||
)
|
||||
assert len(query_result) > 0, (
|
||||
@@ -512,7 +512,7 @@ def updatedoc_partial_check(
|
||||
for k, v in DEFAULT_VECTOR_FIELD_NAME.items():
|
||||
if v != {}:
|
||||
query_result = collection.query(
|
||||
VectorQuery(field_name=v, vector=update_doc_full.vectors[v]),
|
||||
Query(field_name=v, vector=update_doc_full.vectors[v]),
|
||||
topk=10,
|
||||
)
|
||||
assert len(query_result) > 0, (
|
||||
@@ -580,7 +580,7 @@ def batchdoc_and_check(collection, multiple_docs, doc_num, operator="insert"):
|
||||
first_doc = multiple_docs[doc_num - 1]
|
||||
for k, v in DEFAULT_VECTOR_FIELD_NAME.items():
|
||||
query_result = collection.query(
|
||||
VectorQuery(field_name=v, vector=first_doc.vectors[v]),
|
||||
Query(field_name=v, vector=first_doc.vectors[v]),
|
||||
topk=1024,
|
||||
)
|
||||
assert len(query_result) > 0, (
|
||||
|
||||
@@ -19,7 +19,7 @@ from fixture_helper import *
|
||||
from params_helper import *
|
||||
from zvec import StatusCode
|
||||
from zvec.extension import QwenReRanker, RrfReRanker, WeightedReRanker
|
||||
from zvec.model import Collection, Doc, VectorQuery
|
||||
from zvec.model import Collection, Doc
|
||||
from zvec.model.param import (
|
||||
CollectionOption,
|
||||
FlatIndexParam,
|
||||
@@ -81,7 +81,7 @@ def batchdoc_and_check(
|
||||
first_doc = multiple_docs[doc_num - 1]
|
||||
for k, v in DEFAULT_VECTOR_FIELD_NAME.items():
|
||||
query_result = collection.query(
|
||||
VectorQuery(field_name=v, vector=first_doc.vectors[v]),
|
||||
Query(field_name=v, vector=first_doc.vectors[v]),
|
||||
topk=1024,
|
||||
include_vector=True,
|
||||
)
|
||||
@@ -152,7 +152,7 @@ def batchdoc_and_check_ivf(
|
||||
for k, v in DEFAULT_VECTOR_FIELD_NAME.items():
|
||||
if v in ["vector_fp16_field", "vector_fp32_field"]:
|
||||
query_result = collection.query(
|
||||
VectorQuery(field_name=v, vector=first_doc.vectors[v]),
|
||||
Query(field_name=v, vector=first_doc.vectors[v]),
|
||||
topk=1024,
|
||||
include_vector=True,
|
||||
)
|
||||
@@ -571,7 +571,7 @@ class TestCollectionQuery:
|
||||
full_collection_new, multiple_docs, doc_num, operator="insert"
|
||||
)
|
||||
for k, v in DEFAULT_VECTOR_FIELD_NAME.items():
|
||||
query_result = full_collection_new.query(VectorQuery(field_name=v, id="1"))
|
||||
query_result = full_collection_new.query(Query(field_name=v, id="1"))
|
||||
assert len(query_result) > 0
|
||||
query_doc = full_collection_new.fetch(ids=["1"])
|
||||
query_vector = query_doc["1"].vector(v)
|
||||
@@ -595,9 +595,7 @@ class TestCollectionQuery:
|
||||
)
|
||||
for k, v in DEFAULT_VECTOR_FIELD_NAME.items():
|
||||
if v in ["vector_fp16_field", "vector_fp32_field"]:
|
||||
query_result = full_collection_ivf.query(
|
||||
VectorQuery(field_name=v, id="1")
|
||||
)
|
||||
query_result = full_collection_ivf.query(Query(field_name=v, id="1"))
|
||||
assert len(query_result) > 0
|
||||
query_doc = full_collection_ivf.fetch(ids=["1"])
|
||||
query_vector = query_doc["1"].vector(v)
|
||||
@@ -635,22 +633,22 @@ class TestCollectionQuery:
|
||||
query_vector = doc_vectors[v]
|
||||
if topk and filter:
|
||||
query_result = full_collection_new.query(
|
||||
Query(field_name=v, vector=query_vector),
|
||||
filter=filter,
|
||||
vectors=VectorQuery(field_name=v, vector=query_vector),
|
||||
topk=topk,
|
||||
)
|
||||
elif topk and not filter:
|
||||
query_result = full_collection_new.query(
|
||||
VectorQuery(field_name=v, vector=query_vector), topk=topk
|
||||
Query(field_name=v, vector=query_vector), topk=topk
|
||||
)
|
||||
elif not topk and filter:
|
||||
query_result = full_collection_new.query(
|
||||
Query(field_name=v, vector=query_vector),
|
||||
filter=filter,
|
||||
vectors=VectorQuery(field_name=v, vector=query_vector),
|
||||
)
|
||||
else:
|
||||
query_result = full_collection_new.query(
|
||||
VectorQuery(field_name=v, vector=query_vector)
|
||||
Query(field_name=v, vector=query_vector)
|
||||
)
|
||||
assert len(query_result) > 0, (
|
||||
f"Expected at least 1 query result, but got {len(query_result)}"
|
||||
@@ -680,7 +678,7 @@ class TestCollectionQuery:
|
||||
)
|
||||
query_vector = doc_vectors[v]
|
||||
query_result = full_collection_ivf.query(
|
||||
VectorQuery(field_name=v, vector=query_vector),
|
||||
Query(field_name=v, vector=query_vector),
|
||||
topk=1024,
|
||||
)
|
||||
assert len(query_result) > 0, (
|
||||
@@ -707,16 +705,16 @@ class TestCollectionQuery:
|
||||
single_query_results = {}
|
||||
for k, v in DEFAULT_VECTOR_FIELD_NAME.items():
|
||||
single_query_results[v] = full_collection.query(
|
||||
VectorQuery(field_name=v, vector=doc_vectors[v])
|
||||
Query(field_name=v, vector=doc_vectors[v])
|
||||
)
|
||||
expected_rrf_scores = calculate_multi_vector_rrf_scores(single_query_results)
|
||||
multi_query_vectors = []
|
||||
for k, v in DEFAULT_VECTOR_FIELD_NAME.items():
|
||||
multi_query_vectors.append(VectorQuery(field_name=v, vector=doc_vectors[v]))
|
||||
multi_query_vectors.append(Query(field_name=v, vector=doc_vectors[v]))
|
||||
|
||||
rrf_reranker = RrfReRanker(topn=3)
|
||||
multi_query_result = full_collection.query(
|
||||
vectors=multi_query_vectors,
|
||||
multi_query_vectors,
|
||||
reranker=rrf_reranker,
|
||||
)
|
||||
assert len(multi_query_result) > 0, (
|
||||
@@ -773,7 +771,7 @@ class TestCollectionQuery:
|
||||
single_query_results = {}
|
||||
for k, v in DEFAULT_VECTOR_FIELD_NAME.items():
|
||||
single_query_results[v] = full_collection.query(
|
||||
VectorQuery(field_name=v, vector=doc_vectors[v])
|
||||
Query(field_name=v, vector=doc_vectors[v])
|
||||
)
|
||||
expected_weighted_scores = calculate_multi_vector_weighted_scores(
|
||||
single_query_results, weights, MetricType.IP
|
||||
@@ -781,10 +779,10 @@ class TestCollectionQuery:
|
||||
|
||||
multi_query_vectors = []
|
||||
for k, v in DEFAULT_VECTOR_FIELD_NAME.items():
|
||||
multi_query_vectors.append(VectorQuery(field_name=v, vector=doc_vectors[v]))
|
||||
multi_query_vectors.append(Query(field_name=v, vector=doc_vectors[v]))
|
||||
|
||||
multi_query_result = full_collection.query(
|
||||
vectors=multi_query_vectors,
|
||||
multi_query_vectors,
|
||||
reranker=weighted_reranker,
|
||||
)
|
||||
assert len(multi_query_result) > 0, (
|
||||
@@ -868,10 +866,8 @@ class TestCollectionQuery:
|
||||
)
|
||||
query_vector = doc_vectors[v]
|
||||
query_result = full_collection_new.query(
|
||||
Query(field_name=v, vector=query_vector, param=HnswQueryParam(ef=ef)),
|
||||
filter=filter,
|
||||
vectors=VectorQuery(
|
||||
field_name=v, vector=query_vector, param=HnswQueryParam(ef=ef)
|
||||
),
|
||||
topk=topk,
|
||||
)
|
||||
assert len(query_result) > 0, (
|
||||
@@ -903,10 +899,10 @@ class TestCollectionQuery:
|
||||
query_vector = doc_vectors[v]
|
||||
with pytest.raises(Exception) as exc_info:
|
||||
full_collection.query(
|
||||
filter=filter,
|
||||
vectors=VectorQuery(
|
||||
Query(
|
||||
field_name=v, vector=query_vector, param=HnswQueryParam(ef=ef)
|
||||
),
|
||||
filter=filter,
|
||||
topk=topk,
|
||||
)
|
||||
assert INCOMPATIBLE_CONSTRUCTOR_ERROR_MSG in str(exc_info.value)
|
||||
@@ -935,12 +931,12 @@ class TestCollectionQuery:
|
||||
query_vector = doc_vectors[v]
|
||||
|
||||
query_result = full_collection_ivf.query(
|
||||
filter=filter,
|
||||
vectors=VectorQuery(
|
||||
Query(
|
||||
field_name=v,
|
||||
vector=query_vector,
|
||||
param=IVFQueryParam(nprobe=nprobe),
|
||||
),
|
||||
filter=filter,
|
||||
topk=topk,
|
||||
)
|
||||
assert len(query_result) > 0
|
||||
@@ -977,12 +973,12 @@ class TestCollectionQuery:
|
||||
query_vector = doc_vectors[v]
|
||||
with pytest.raises(Exception) as exc_info:
|
||||
full_collection_ivf.query(
|
||||
# filter=filter,
|
||||
vectors=VectorQuery(
|
||||
Query(
|
||||
field_name=v,
|
||||
vector=query_vector,
|
||||
param=IVFQueryParam(nprobe=nprobe),
|
||||
),
|
||||
# filter=filter,
|
||||
topk=topk,
|
||||
)
|
||||
assert INCOMPATIBLE_CONSTRUCTOR_ERROR_MSG in str(exc_info.value)
|
||||
@@ -1005,10 +1001,10 @@ class TestCollectionQuery:
|
||||
query_vector = doc_vectors[v]
|
||||
if v in ["vector_fp16_field", "vector_fp32_field"]:
|
||||
full_collection.query(
|
||||
filter=filter,
|
||||
vectors=VectorQuery(
|
||||
Query(
|
||||
field_name=v, vector=query_vector, param=HnswIndexParam()
|
||||
),
|
||||
filter=filter,
|
||||
)
|
||||
assert INCOMPATIBLE_FUNCTION_ERROR_MSG in str(exc_info.value)
|
||||
|
||||
@@ -1018,56 +1014,56 @@ class TestCollectionQuery:
|
||||
[
|
||||
(
|
||||
"Non-existent vector field name",
|
||||
lambda ref_dense_vector: VectorQuery(
|
||||
lambda ref_dense_vector: Query(
|
||||
field_name="nonexistent_vector", vector=ref_dense_vector
|
||||
),
|
||||
"Expected exception for non-existent vector field name",
|
||||
),
|
||||
(
|
||||
"Invalid vector data type for dense vector (string instead of list)",
|
||||
lambda ref_dense_vector: VectorQuery(
|
||||
lambda ref_dense_vector: Query(
|
||||
field_name="vector_fp32_field", vector="invalid_vector_data"
|
||||
),
|
||||
"Expected exception for invalid dense vector data type",
|
||||
),
|
||||
(
|
||||
"Invalid vector data type for sparse vector (list instead of dict)",
|
||||
lambda ref_dense_vector: VectorQuery(
|
||||
lambda ref_dense_vector: Query(
|
||||
field_name="sparse_fp32", vector=[1.0, 2.0, 3.0]
|
||||
),
|
||||
"Expected exception for invalid sparse vector data type",
|
||||
),
|
||||
(
|
||||
"Empty vector data for dense vector",
|
||||
lambda ref_dense_vector: VectorQuery(
|
||||
lambda ref_dense_vector: Query(
|
||||
field_name="vector_fp32_field", vector=[]
|
||||
),
|
||||
"Expected exception for empty dense vector data",
|
||||
),
|
||||
(
|
||||
"Invalid dimension for dense vector",
|
||||
lambda ref_dense_vector: VectorQuery(
|
||||
lambda ref_dense_vector: Query(
|
||||
field_name="vector_fp32_field", vector=[1.0, 2.0]
|
||||
), # Only 2 dimensions instead of 128
|
||||
"Expected exception for invalid dense vector dimension",
|
||||
),
|
||||
(
|
||||
"Non-existent document ID for by_id query",
|
||||
lambda ref_dense_vector: VectorQuery(
|
||||
lambda ref_dense_vector: Query(
|
||||
field_name="vector_fp32_field", id="999"
|
||||
), # Non-existent ID
|
||||
"Expected exception for non-existent document ID",
|
||||
),
|
||||
(
|
||||
"Both vector and id specified (invalid combination)",
|
||||
lambda ref_dense_vector: VectorQuery(
|
||||
lambda ref_dense_vector: Query(
|
||||
field_name="vector_fp32_field", vector=ref_dense_vector, id="5"
|
||||
),
|
||||
"Expected exception for specifying both vector and id",
|
||||
),
|
||||
(
|
||||
"Neither vector nor id specified",
|
||||
lambda ref_dense_vector: VectorQuery(
|
||||
lambda ref_dense_vector: Query(
|
||||
field_name="vector_fp32_field"
|
||||
), # Neither vector nor id
|
||||
"Expected exception for specifying neither vector nor id",
|
||||
@@ -1092,7 +1088,7 @@ class TestCollectionQuery:
|
||||
ref_dense_vector = ref_doc.vector("vector_fp32_field")
|
||||
|
||||
with pytest.raises(Exception) as exc_info:
|
||||
full_collection.query(vectors=[vector_query(ref_dense_vector)])
|
||||
full_collection.query([vector_query(ref_dense_vector)])
|
||||
assert exc_info.value is not None, expected_error_msg
|
||||
|
||||
@pytest.mark.parametrize("filter", ["int32_field >= 3 and int32_field <= 7"])
|
||||
@@ -1112,8 +1108,8 @@ class TestCollectionQuery:
|
||||
)
|
||||
query_vector = doc_vectors[v]
|
||||
full_collection.query(
|
||||
Query(field_name=v, vector=query_vector),
|
||||
filter=filter,
|
||||
vectors=VectorQuery(field_name=v, vector=query_vector),
|
||||
param=HnswIndexParam(),
|
||||
topk=3,
|
||||
)
|
||||
@@ -1262,9 +1258,7 @@ class TestCollectionConcurrencyOperations:
|
||||
else:
|
||||
query_vector = [0.1] * 128
|
||||
result = full_collection.query(
|
||||
VectorQuery(
|
||||
field_name="vector_fp32_field", vector=query_vector
|
||||
),
|
||||
Query(field_name="vector_fp32_field", vector=query_vector),
|
||||
topk=3,
|
||||
)
|
||||
|
||||
|
||||
@@ -26,8 +26,8 @@ from zvec import (
|
||||
Collection,
|
||||
Doc,
|
||||
FieldSchema,
|
||||
Query,
|
||||
VectorSchema,
|
||||
VectorQuery,
|
||||
)
|
||||
|
||||
|
||||
@@ -182,11 +182,11 @@ class TestCollectionExceptionHandling:
|
||||
"Expected exception for missing ids parameter"
|
||||
)
|
||||
|
||||
def test_query_missing_vectorquery_field_name(self, test_collection: Collection):
|
||||
def test_query_missing_query_field_name(self, test_collection: Collection):
|
||||
with pytest.raises(Exception) as exc_info:
|
||||
result = test_collection.query(vectors=[VectorQuery()])
|
||||
result = test_collection.query([Query()])
|
||||
assert exc_info.value is not None, (
|
||||
"Expected exception for missing VectorQuery field_name parameter"
|
||||
"Expected exception for missing Query field_name parameter"
|
||||
)
|
||||
|
||||
def test_add_column_missing_field_schema(self, test_collection: Collection):
|
||||
|
||||
@@ -478,7 +478,7 @@ class TestCollectionOpen:
|
||||
# Test vector query functionality for dense vectors
|
||||
query_vector_dense = [0.1] * 128
|
||||
vector_query_result = opened_coll.query(
|
||||
VectorQuery(field_name="dense", vector=query_vector_dense)
|
||||
Query(field_name="dense", vector=query_vector_dense)
|
||||
)
|
||||
assert len(vector_query_result) > 0, (
|
||||
f"Expected at least 1 vector query result, but got {len(vector_query_result)}"
|
||||
@@ -529,7 +529,7 @@ class TestCollectionOpen:
|
||||
# Test vector query functionality for sparse vectors
|
||||
query_vector_sparse = {1: 1.0, 2: 2.0, 3: 3.0}
|
||||
sparse_vector_query_result = opened_coll.query(
|
||||
VectorQuery(field_name="sparse", vector=query_vector_sparse)
|
||||
Query(field_name="sparse", vector=query_vector_sparse)
|
||||
)
|
||||
assert len(sparse_vector_query_result) > 0, (
|
||||
f"Expected at least 1 sparse vector query result, but got {len(sparse_vector_query_result)}"
|
||||
|
||||
@@ -15,7 +15,7 @@
|
||||
import pytest
|
||||
|
||||
from zvec.typing import DataType, StatusCode, MetricType, QuantizeType
|
||||
from zvec.model import Collection, Doc, VectorQuery
|
||||
from zvec.model import Collection, Doc, Query
|
||||
from zvec.model.param import (
|
||||
CollectionOption,
|
||||
InvertIndexParam,
|
||||
@@ -158,7 +158,7 @@ def get_ground_truth_for_vector_query(
|
||||
|
||||
else:
|
||||
full_result = collection.query(
|
||||
VectorQuery(field_name=field_name, vector=query_vector),
|
||||
Query(field_name=field_name, vector=query_vector),
|
||||
topk=min(len(all_docs), 1024),
|
||||
include_vector=True,
|
||||
)
|
||||
@@ -214,7 +214,7 @@ def calculate_recall_at_k(
|
||||
print("Starting %dth query" % i)
|
||||
|
||||
query_result_list = collection.query(
|
||||
VectorQuery(field_name=field_name, vector=query_vector),
|
||||
Query(field_name=field_name, vector=query_vector),
|
||||
topk=1024,
|
||||
include_vector=True,
|
||||
)
|
||||
|
||||
@@ -29,7 +29,7 @@ from zvec import (
|
||||
LogType,
|
||||
OptimizeOption,
|
||||
StatusCode,
|
||||
VectorQuery,
|
||||
Query,
|
||||
VectorSchema,
|
||||
)
|
||||
|
||||
@@ -934,7 +934,7 @@ class TestCollectionQuery:
|
||||
):
|
||||
with pytest.raises(ValueError):
|
||||
collection_with_single_doc.query(
|
||||
VectorQuery(
|
||||
Query(
|
||||
field_name="dense",
|
||||
id=single_doc.id,
|
||||
vector=single_doc.vector("dense"),
|
||||
@@ -950,19 +950,19 @@ class TestCollectionQuery:
|
||||
def test_collection_with_error_query_vector(
|
||||
self, collection_with_multiple_docs: Collection, multiple_docs
|
||||
):
|
||||
query = VectorQuery(
|
||||
query = Query(
|
||||
field_name="dense", vector=multiple_docs[0].vector("dense"), param=[1, 2, 3]
|
||||
)
|
||||
with pytest.raises(TypeError):
|
||||
result = collection_with_multiple_docs.query(
|
||||
filter="id in (1)", topk=100, vectors=query
|
||||
query, filter="id in (1)", topk=100
|
||||
)
|
||||
|
||||
def test_collection_query_by_id(
|
||||
self, collection_with_multiple_docs: Collection, multiple_docs
|
||||
):
|
||||
result = collection_with_multiple_docs.query(
|
||||
VectorQuery(field_name="dense", id=multiple_docs[0].id)
|
||||
Query(field_name="dense", id=multiple_docs[0].id)
|
||||
)
|
||||
assert len(result) == 10
|
||||
|
||||
@@ -972,12 +972,8 @@ class TestCollectionQuery:
|
||||
with pytest.raises(ValueError):
|
||||
collection_with_multiple_docs.query(
|
||||
[
|
||||
VectorQuery(
|
||||
field_name="dense", vector=multiple_docs[0].vector("dense")
|
||||
),
|
||||
VectorQuery(
|
||||
field_name="dense", vector=multiple_docs[0].vector("dense")
|
||||
),
|
||||
Query(field_name="dense", vector=multiple_docs[0].vector("dense")),
|
||||
Query(field_name="dense", vector=multiple_docs[0].vector("dense")),
|
||||
]
|
||||
)
|
||||
|
||||
|
||||
@@ -34,7 +34,7 @@ from zvec import (
|
||||
HnswRabitqQueryParam,
|
||||
MetricType,
|
||||
VectorSchema,
|
||||
VectorQuery,
|
||||
Query,
|
||||
)
|
||||
|
||||
|
||||
@@ -285,13 +285,13 @@ class TestHnswRabitqCollectionQuery:
|
||||
):
|
||||
"""Test querying by vector with HNSW RaBitQ index."""
|
||||
query_vector = multiple_docs[0].vector("embedding")
|
||||
query = VectorQuery(
|
||||
query = Query(
|
||||
field_name="embedding",
|
||||
vector=query_vector,
|
||||
param=HnswRabitqQueryParam(ef=300),
|
||||
)
|
||||
|
||||
result = collection_with_multiple_docs.query(vectors=query, topk=10)
|
||||
result = collection_with_multiple_docs.query(queries=query, topk=10)
|
||||
assert len(result) > 0
|
||||
assert len(result) <= 10
|
||||
|
||||
@@ -304,13 +304,13 @@ class TestHnswRabitqCollectionQuery:
|
||||
self, collection_with_multiple_docs: Collection, multiple_docs: list[Doc]
|
||||
):
|
||||
"""Test querying by document ID with HNSW RaBitQ index."""
|
||||
query = VectorQuery(
|
||||
query = Query(
|
||||
field_name="embedding",
|
||||
id=multiple_docs[0].id,
|
||||
param=HnswRabitqQueryParam(ef=300),
|
||||
)
|
||||
|
||||
result = collection_with_multiple_docs.query(vectors=query, topk=10)
|
||||
result = collection_with_multiple_docs.query(queries=query, topk=10)
|
||||
assert len(result) > 0
|
||||
assert len(result) <= 10
|
||||
|
||||
@@ -321,21 +321,21 @@ class TestHnswRabitqCollectionQuery:
|
||||
query_vector = multiple_docs[0].vector("embedding")
|
||||
|
||||
# Test with ef=100
|
||||
query_100 = VectorQuery(
|
||||
query_100 = Query(
|
||||
field_name="embedding",
|
||||
vector=query_vector,
|
||||
param=HnswRabitqQueryParam(ef=100),
|
||||
)
|
||||
result_100 = collection_with_multiple_docs.query(vectors=query_100, topk=10)
|
||||
result_100 = collection_with_multiple_docs.query(queries=query_100, topk=10)
|
||||
assert len(result_100) > 0
|
||||
|
||||
# Test with ef=500
|
||||
query_500 = VectorQuery(
|
||||
query_500 = Query(
|
||||
field_name="embedding",
|
||||
vector=query_vector,
|
||||
param=HnswRabitqQueryParam(ef=500),
|
||||
)
|
||||
result_500 = collection_with_multiple_docs.query(vectors=query_500, topk=10)
|
||||
result_500 = collection_with_multiple_docs.query(queries=query_500, topk=10)
|
||||
assert len(result_500) > 0
|
||||
|
||||
def test_query_with_topk(
|
||||
@@ -343,18 +343,18 @@ class TestHnswRabitqCollectionQuery:
|
||||
):
|
||||
"""Test querying with different topk values."""
|
||||
query_vector = multiple_docs[0].vector("embedding")
|
||||
query = VectorQuery(
|
||||
query = Query(
|
||||
field_name="embedding",
|
||||
vector=query_vector,
|
||||
param=HnswRabitqQueryParam(ef=300),
|
||||
)
|
||||
|
||||
# Test topk=5
|
||||
result_5 = collection_with_multiple_docs.query(vectors=query, topk=5)
|
||||
result_5 = collection_with_multiple_docs.query(queries=query, topk=5)
|
||||
assert len(result_5) <= 5
|
||||
|
||||
# Test topk=20
|
||||
result_20 = collection_with_multiple_docs.query(vectors=query, topk=20)
|
||||
result_20 = collection_with_multiple_docs.query(queries=query, topk=20)
|
||||
assert len(result_20) <= 20
|
||||
|
||||
def test_query_with_filter(
|
||||
@@ -362,7 +362,7 @@ class TestHnswRabitqCollectionQuery:
|
||||
):
|
||||
"""Test querying with filter conditions."""
|
||||
query_vector = multiple_docs[0].vector("embedding")
|
||||
query = VectorQuery(
|
||||
query = Query(
|
||||
field_name="embedding",
|
||||
vector=query_vector,
|
||||
param=HnswRabitqQueryParam(ef=300),
|
||||
@@ -370,7 +370,7 @@ class TestHnswRabitqCollectionQuery:
|
||||
|
||||
# Query with id filter
|
||||
result = collection_with_multiple_docs.query(
|
||||
vectors=query, topk=10, filter="id < 50"
|
||||
queries=query, topk=10, filter="id < 50"
|
||||
)
|
||||
assert len(result) > 0
|
||||
for doc in result:
|
||||
@@ -381,14 +381,14 @@ class TestHnswRabitqCollectionQuery:
|
||||
):
|
||||
"""Test querying with specific output fields."""
|
||||
query_vector = multiple_docs[0].vector("embedding")
|
||||
query = VectorQuery(
|
||||
query = Query(
|
||||
field_name="embedding",
|
||||
vector=query_vector,
|
||||
param=HnswRabitqQueryParam(ef=300),
|
||||
)
|
||||
|
||||
result = collection_with_multiple_docs.query(
|
||||
vectors=query, topk=10, output_fields=["id", "name"]
|
||||
queries=query, topk=10, output_fields=["id", "name"]
|
||||
)
|
||||
assert len(result) > 0
|
||||
|
||||
@@ -401,14 +401,14 @@ class TestHnswRabitqCollectionQuery:
|
||||
):
|
||||
"""Test querying with vector data included in results."""
|
||||
query_vector = multiple_docs[0].vector("embedding")
|
||||
query = VectorQuery(
|
||||
query = Query(
|
||||
field_name="embedding",
|
||||
vector=query_vector,
|
||||
param=HnswRabitqQueryParam(ef=300),
|
||||
)
|
||||
|
||||
result = collection_with_multiple_docs.query(
|
||||
vectors=query, topk=10, include_vector=True
|
||||
queries=query, topk=10, include_vector=True
|
||||
)
|
||||
assert len(result) > 0
|
||||
|
||||
@@ -536,12 +536,12 @@ class TestHnswRabitqCollectionOptimizeAndReopen:
|
||||
|
||||
# Verify data is still accessible after optimize
|
||||
query_vector = multiple_docs[0].vector("embedding")
|
||||
query = VectorQuery(
|
||||
query = Query(
|
||||
field_name="embedding",
|
||||
vector=query_vector,
|
||||
param=HnswRabitqQueryParam(ef=300),
|
||||
)
|
||||
result_before_close = coll.query(vectors=query, topk=10)
|
||||
result_before_close = coll.query(query, topk=10)
|
||||
assert len(result_before_close) > 0
|
||||
|
||||
# Close collection (destroy will close it)
|
||||
@@ -554,12 +554,12 @@ class TestHnswRabitqCollectionOptimizeAndReopen:
|
||||
assert reopened_coll.stats.doc_count == len(multiple_docs)
|
||||
|
||||
# Execute query on reopened collection
|
||||
query_after_reopen = VectorQuery(
|
||||
query_after_reopen = Query(
|
||||
field_name="embedding",
|
||||
vector=query_vector,
|
||||
param=HnswRabitqQueryParam(ef=300),
|
||||
)
|
||||
result_after_reopen = reopened_coll.query(vectors=query_after_reopen, topk=10)
|
||||
result_after_reopen = reopened_coll.query(query_after_reopen, topk=10)
|
||||
assert len(result_after_reopen) > 0
|
||||
assert len(result_after_reopen) <= 10
|
||||
|
||||
|
||||
@@ -31,6 +31,7 @@ from zvec import (
|
||||
OptimizeOption,
|
||||
HnswQueryParam,
|
||||
IVFQueryParam,
|
||||
Query,
|
||||
VectorQuery,
|
||||
IndexType,
|
||||
MetricType,
|
||||
@@ -373,9 +374,9 @@ class TestHnswQueryParam:
|
||||
# assert param.scale_factor == 40
|
||||
|
||||
|
||||
class TestVectorQuery:
|
||||
class TestQuery:
|
||||
def test_init_with_valid_id(self):
|
||||
vq = VectorQuery(field_name="embedding", id="doc123")
|
||||
vq = Query(field_name="embedding", id="doc123")
|
||||
assert vq.field_name == "embedding"
|
||||
assert vq.id == "doc123"
|
||||
assert vq.vector is None
|
||||
@@ -384,32 +385,52 @@ class TestVectorQuery:
|
||||
def test_init_with_valid_vector(self):
|
||||
vec = [0.1, 0.2, 0.3]
|
||||
param = HnswQueryParam(ef=300)
|
||||
vq = VectorQuery(field_name="embedding", vector=vec, param=param)
|
||||
vq = Query(field_name="embedding", vector=vec, param=param)
|
||||
assert vq.field_name == "embedding"
|
||||
assert vq.vector == vec
|
||||
assert vq.param == param
|
||||
|
||||
def test_init_both_id_and_vector_raises_error(self):
|
||||
with pytest.raises(ValueError):
|
||||
VectorQuery(field_name="embedding", id="doc123", vector=[0.1])._validate()
|
||||
Query(field_name="embedding", id="doc123", vector=[0.1])._validate()
|
||||
|
||||
def test_init_without_field_name_raises_error(self):
|
||||
with pytest.raises(ValueError):
|
||||
VectorQuery(field_name=None)._validate()
|
||||
Query(field_name=None)._validate()
|
||||
|
||||
def test_has_id_returns_true_when_id_set(self):
|
||||
vq = VectorQuery(field_name="embedding", id="doc123")
|
||||
vq = Query(field_name="embedding", id="doc123")
|
||||
assert vq.has_id()
|
||||
|
||||
def test_has_id_returns_false_when_no_id(self):
|
||||
vq = VectorQuery(field_name="embedding", vector=[0.1])
|
||||
vq = Query(field_name="embedding", vector=[0.1])
|
||||
assert not vq.has_id()
|
||||
|
||||
def test_has_vector_returns_true_with_non_empty_vector(self):
|
||||
vq = VectorQuery(field_name="embedding", vector=[0.1])
|
||||
vq = Query(field_name="embedding", vector=[0.1])
|
||||
assert vq.has_vector()
|
||||
|
||||
def test_validate_fails_on_both_id_and_vector(self):
|
||||
vq = VectorQuery(field_name="test", id="doc123", vector=[0.1])
|
||||
vq = Query(field_name="test", id="doc123", vector=[0.1])
|
||||
with pytest.raises(ValueError):
|
||||
vq._validate()
|
||||
|
||||
|
||||
class TestVectorQueryDeprecated:
|
||||
def test_deprecation_warning(self):
|
||||
import warnings
|
||||
|
||||
with warnings.catch_warnings(record=True) as w:
|
||||
warnings.simplefilter("always")
|
||||
vq = VectorQuery(field_name="embedding", id="doc123")
|
||||
assert len(w) == 1
|
||||
assert issubclass(w[0].category, DeprecationWarning)
|
||||
assert "Query" in str(w[0].message)
|
||||
|
||||
def test_isinstance_compatibility(self):
|
||||
import warnings
|
||||
|
||||
with warnings.catch_warnings(record=True):
|
||||
warnings.simplefilter("always")
|
||||
vq = VectorQuery(field_name="embedding", id="doc123")
|
||||
assert isinstance(vq, Query)
|
||||
|
||||
@@ -28,9 +28,16 @@ from zvec.executor.query_executor import (
|
||||
QueryExecutor,
|
||||
QueryExecutorFactory,
|
||||
SingleVectorQueryExecutor,
|
||||
)
|
||||
from zvec import (
|
||||
RrfReRanker,
|
||||
HnswQueryParam,
|
||||
CollectionSchema,
|
||||
VectorSchema,
|
||||
DataType,
|
||||
Query,
|
||||
VectorQuery,
|
||||
)
|
||||
from zvec import RrfReRanker, HnswQueryParam, CollectionSchema, VectorSchema, DataType
|
||||
|
||||
|
||||
# ----------------------------
|
||||
@@ -65,16 +72,16 @@ class MockCollectionSchema(CollectionSchema):
|
||||
# ----------------------------
|
||||
# VectorQuery Test Case
|
||||
# ----------------------------
|
||||
class TestVectorQuery:
|
||||
class TestQuery:
|
||||
def test_init(self):
|
||||
query = VectorQuery(field_name="test_field")
|
||||
query = Query(field_name="test_field")
|
||||
assert query.field_name == "test_field"
|
||||
assert query.id is None
|
||||
assert query.vector is None
|
||||
assert query.param is None
|
||||
|
||||
param = HnswQueryParam()
|
||||
query = VectorQuery(
|
||||
query = Query(
|
||||
field_name="test_field", id="test_id", vector=[1, 2, 3], param=param
|
||||
)
|
||||
assert query.field_name == "test_field"
|
||||
@@ -83,20 +90,20 @@ class TestVectorQuery:
|
||||
assert query.param == param
|
||||
|
||||
def test_has_id(self):
|
||||
query = VectorQuery(field_name="test_field")
|
||||
query = Query(field_name="test_field")
|
||||
assert not query.has_id()
|
||||
|
||||
query = VectorQuery(field_name="test_field", id="test_id")
|
||||
query = Query(field_name="test_field", id="test_id")
|
||||
assert query.has_id()
|
||||
|
||||
def test_has_vector(self):
|
||||
query = VectorQuery(field_name="test_field")
|
||||
query = Query(field_name="test_field")
|
||||
assert not query.has_vector()
|
||||
|
||||
query = VectorQuery(field_name="test_field", vector=[])
|
||||
query = Query(field_name="test_field", vector=[])
|
||||
assert not query.has_vector()
|
||||
|
||||
query = VectorQuery(field_name="test_field", vector=[1, 2, 3])
|
||||
query = Query(field_name="test_field", vector=[1, 2, 3])
|
||||
assert query.has_vector()
|
||||
|
||||
def test_validate_dense_fp16_convert(self):
|
||||
@@ -150,6 +157,26 @@ class TestVectorQuery:
|
||||
assert math.isclose(np.float16(vec[k]), ret[k], abs_tol=1e-6)
|
||||
|
||||
|
||||
class TestVectorQueryDeprecated:
|
||||
def test_deprecation_warning(self):
|
||||
import warnings
|
||||
|
||||
with warnings.catch_warnings(record=True) as w:
|
||||
warnings.simplefilter("always")
|
||||
vq = VectorQuery(field_name="test_field")
|
||||
assert len(w) == 1
|
||||
assert issubclass(w[0].category, DeprecationWarning)
|
||||
assert "Query" in str(w[0].message)
|
||||
|
||||
def test_isinstance_compatibility(self):
|
||||
import warnings
|
||||
|
||||
with warnings.catch_warnings(record=True):
|
||||
warnings.simplefilter("always")
|
||||
vq = VectorQuery(field_name="test_field")
|
||||
assert isinstance(vq, Query)
|
||||
|
||||
|
||||
class TestQueryContext:
|
||||
def test_init(self):
|
||||
ctx = QueryContext(topk=10)
|
||||
@@ -162,7 +189,7 @@ class TestQueryContext:
|
||||
assert ctx.core_vectors == []
|
||||
|
||||
def test_properties(self):
|
||||
queries = [VectorQuery(field_name="test")]
|
||||
queries = [Query(field_name="test")]
|
||||
reranker = RrfReRanker()
|
||||
output_fields = ["field1", "field2"]
|
||||
|
||||
@@ -198,7 +225,7 @@ class TestNoVectorQueryExecutor:
|
||||
def test_do_validate_with_queries(self):
|
||||
schema = MockCollectionSchema()
|
||||
executor = NoVectorQueryExecutor(schema)
|
||||
ctx = QueryContext(topk=10, queries=[VectorQuery(field_name="test")])
|
||||
ctx = QueryContext(topk=10, queries=[Query(field_name="test")])
|
||||
|
||||
with pytest.raises(
|
||||
ValueError, match="Collection does not support query with vector or id"
|
||||
@@ -232,7 +259,7 @@ class TestSingleVectorQueryExecutor:
|
||||
def test_do_validate_multiple_queries(self):
|
||||
schema = MockCollectionSchema()
|
||||
executor = SingleVectorQueryExecutor(schema)
|
||||
queries = [VectorQuery(field_name="test1"), VectorQuery(field_name="test2")]
|
||||
queries = [Query(field_name="test1"), Query(field_name="test2")]
|
||||
ctx = QueryContext(topk=10, queries=queries)
|
||||
|
||||
with pytest.raises(
|
||||
@@ -260,7 +287,7 @@ class TestMultiVectorQueryExecutor:
|
||||
def test_do_validate_multiple_queries_without_reranker(self):
|
||||
schema = MockCollectionSchema()
|
||||
executor = MultiVectorQueryExecutor(schema)
|
||||
queries = [VectorQuery(field_name="test1"), VectorQuery(field_name="test2")]
|
||||
queries = [Query(field_name="test1"), Query(field_name="test2")]
|
||||
ctx = QueryContext(topk=10, queries=queries)
|
||||
|
||||
with pytest.raises(
|
||||
@@ -271,7 +298,7 @@ class TestMultiVectorQueryExecutor:
|
||||
def test_do_validate_multiple_queries_with_reranker(self):
|
||||
schema = MockCollectionSchema()
|
||||
executor = MultiVectorQueryExecutor(schema)
|
||||
queries = [VectorQuery(field_name="test1"), VectorQuery(field_name="test2")]
|
||||
queries = [Query(field_name="test1"), Query(field_name="test2")]
|
||||
reranker = RrfReRanker()
|
||||
ctx = QueryContext(topk=10, queries=queries, reranker=reranker)
|
||||
|
||||
|
||||
Reference in New Issue
Block a user