milvus-io
/
milvus
mirror of https://github.com/milvus-io/milvus.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

test: add some mmap cases (#31211) 

issue : https://github.com/milvus-io/milvus/issues/30940

Signed-off-by: elstic <hao.wang@zilliz.com>
pull/31299/head
elstic 2024-03-15 15:25:05 +08:00 committed by GitHub
parent 9e7ce0a8f7
commit 93bb04f884
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 219 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
tests/python_client/testcases/test_collection.py
View File

@ -4533,7 +4533,6 @@ class TestCollectionMmap(TestcaseBase):
assert pro["mmap.enabled"] == 'True'
@pytest.mark.tags(CaseLabel.L1)
@pytest.mark.xfail(reason="issue: #30800")
def test_load_mmap_collection(self):
"""
target: after loading, enable mmap for the collection
@ -4542,17 +4541,17 @@ class TestCollectionMmap(TestcaseBase):
3. enable mmap on collection
expected: raise exception
"""
self._connect()
c_name = cf.gen_unique_str(prefix)
collection_w, _ = self.collection_wrap.init_collection(c_name, schema=default_schema)
collection_w = self.init_collection_wrap(c_name, schema=default_schema)
collection_w.insert(cf.gen_default_list_data())
collection_w.create_index(ct.default_float_vec_field_name, index_params=ct.default_flat_index,
index_name=ct.default_index_name)
collection_w.set_properties({'mmap.enabled': True})
pro = collection_w.describe().get("properties")
pro = collection_w.describe()[0].get("properties")
assert pro["mmap.enabled"] == 'True'
collection_w.load()
collection_w.set_properties({'mmap.enabled': True},
check_task=CheckTasks.err_res,
check_items={ct.err_code: 104,
ct.err_msg: f"collection already loaded"})

77
tests/python_client/testcases/test_index.py
View File

@ -1147,6 +1147,50 @@ class TestNewIndexBase(TestcaseBase):
check_items={"nq": default_nq,
"limit": default_limit})
@pytest.mark.tags(CaseLabel.L2)
def test_rebuild_mmap_index(self):
"""
target: reconstructing an index after an alter index
method: reconstructing an index after an alter index
expected: build indexes normally , index contains mmap information
"""
self._connect()
c_name = cf.gen_unique_str(prefix)
collection_w, _ = self.collection_wrap.init_collection(c_name, schema=default_schema)
collection_w.create_index(ct.default_float_vec_field_name, default_index_params,
index_name=ct.default_index_name)
collection_w.set_properties({'mmap.enabled': True})
pro = collection_w.describe().get("properties")
assert pro["mmap.enabled"] == 'True'
collection_w.alter_index(ct.default_index_name, {'mmap.enabled': True})
assert collection_w.index().params["mmap.enabled"] == 'True'
collection_w.insert(cf.gen_default_list_data())
collection_w.flush()
# check if mmap works after rebuild index
collection_w.create_index(ct.default_float_vec_field_name, default_index_params,
index_name=ct.default_index_name)
assert collection_w.index().params["mmap.enabled"] == 'True'
collection_w.load()
collection_w.release()
# check if mmap works after reloading and rebuilding index.
collection_w.create_index(ct.default_float_vec_field_name, default_index_params,
index_name=ct.default_index_name)
assert collection_w.index().params["mmap.enabled"] == 'True'
pro = collection_w.describe().get("properties")
assert pro["mmap.enabled"] == 'True'
collection_w.load()
vectors = [[random.random() for _ in range(default_dim)] for _ in range(default_nq)]
collection_w.search(vectors[:default_nq], default_search_field,
default_search_params, default_limit,
default_search_exp,
check_task=CheckTasks.check_search_results,
check_items={"nq": default_nq,
"limit": default_limit})
@pytest.mark.tags(CaseLabel.GPU)
class TestNewIndexBinary(TestcaseBase):
@ -1573,6 +1617,39 @@ class TestIndexInvalid(TestcaseBase):
ct.err_msg: f"index type {scalar_index[i]} does not support mmap"})
collection_w.drop_index(index_name)
@pytest.mark.tags(CaseLabel.L2)
def test_alter_index_invalid(self):
"""
target: the alter index of the error parameter
method: the alter index of the error parameter
expected: raise exception
"""
c_name = cf.gen_unique_str(prefix)
collection_w = self.init_collection_wrap(c_name, schema=default_schema)
collection_w.insert(cf.gen_default_list_data())
collection_w.create_index(ct.default_float_vec_field_name, default_index_params,
index_name=ct.default_index_name)
collection_w.alter_index(ct.default_index_name, {'mmap.enabled': "error_value"},
check_task=CheckTasks.err_res,
check_items={ct.err_code: 65535,
ct.err_msg: f"invalid mmap.enabled value"})
collection_w.alter_index(ct.default_index_name, {"error_param_key": 123},
check_task=CheckTasks.err_res,
check_items={ct.err_code: 1100,
ct.err_msg: f"error_param is not configable index param"})
collection_w.alter_index(ct.default_index_name, ["error_param_type"],
check_task=CheckTasks.err_res,
check_items={ct.err_code: 1,
ct.err_msg: f"Unexpected error"})
collection_w.alter_index(ct.default_index_name, None,
check_task=CheckTasks.err_res,
check_items={ct.err_code: 1,
ct.err_msg: f"extra_params should not be None"})
collection_w.alter_index(ct.default_index_name, 1000,
check_task=CheckTasks.err_res,
check_items={ct.err_code: 1,
ct.err_msg: f"<'int' object has no attribute 'items'"})
@pytest.mark.tags(CaseLabel.GPU)
class TestNewIndexAsync(TestcaseBase):

136
tests/python_client/testcases/test_query.py
View File

@ -1985,6 +1985,142 @@ class TestQueryParams(TestcaseBase):
assert [res1[i][default_float_field_name] for i in range(upsert_nb)] == \
[res2[i][default_float_field_name] for i in range(upsert_nb)]
@pytest.mark.tags(CaseLabel.L1)
def test_mmap_query_expr_empty_pk_string(self):
"""
target: turn on mmap to test queries using empty expression
method: enable mmap to query for empty expressions with restrictions.
expected: return the first K results in order
"""
# 1. initialize with data
collection_w, _, _, insert_ids = \
self.init_collection_general(prefix, True, is_index=False, primary_field=ct.default_string_field_name)[0:4]
collection_w.set_properties({'mmap.enabled': True})
# string field is sorted by lexicographical order
exp_ids, res = ['0', '1', '10', '100', '1000', '1001', '1002', '1003', '1004', '1005'], []
for ids in exp_ids:
res.append({ct.default_string_field_name: ids})
collection_w.create_index(ct.default_float_vec_field_name, default_index_params, index_name="query_index")
collection_w.load()
# 2. query with limit
collection_w.query("", limit=ct.default_limit,
check_task=CheckTasks.check_query_results, check_items={exp_res: res})
# 3. query with limit + offset
res = res[5:]
collection_w.query("", limit=5, offset=5,
check_task=CheckTasks.check_query_results, check_items={exp_res: res})
@pytest.mark.tags(CaseLabel.L1)
def test_enable_mmap_query_with_expression(self, get_normal_expr, enable_dynamic_field):
"""
target: turn on mmap use different expr queries
method: turn on mmap and query with different expr
expected: verify query result
"""
# 1. initialize with data
nb = 1000
collection_w, _vectors, _, insert_ids = self.init_collection_general(prefix, True, nb, is_index=False,
enable_dynamic_field=enable_dynamic_field)[0:4]
# enable mmap
collection_w.set_properties({'mmap.enabled': True})
collection_w.create_index(ct.default_float_vec_field_name, default_index_params, index_name="query_expr_index")
collection_w.alter_index("query_expr_index", {'mmap.enabled': True})
collection_w.load()
# filter result with expression in collection
_vectors = _vectors[0]
expr = get_normal_expr
expression = expr.replace("&&", "and").replace("||", "or")
filter_ids = []
for i, _id in enumerate(insert_ids):
if enable_dynamic_field:
int64 = _vectors[i][ct.default_int64_field_name]
float = _vectors[i][ct.default_float_field_name]
else:
int64 = _vectors.int64[i]
float = _vectors.float[i]
if not expression or eval(expression):
filter_ids.append(_id)
# query and verify result
res = collection_w.query(expr=expression)[0]
query_ids = set(map(lambda x: x[ct.default_int64_field_name], res))
assert query_ids == set(filter_ids)
@pytest.mark.tags(CaseLabel.L2)
def test_mmap_query_string_field_not_primary_is_empty(self):
"""
target: enable mmap, use string expr to test query, string field is not the main field
method: create collection , string field is primary
enable mmap
collection load and insert empty data with string field
collection query uses string expr in string field
expected: query successfully
"""
# 1. create a collection
collection_w, vectors = self.init_collection_general(prefix, insert_data=False, is_index=False)[0:2]
nb = 3000
df = cf.gen_default_list_data(nb)
df[2] = ["" for _ in range(nb)]
collection_w.insert(df)
assert collection_w.num_entities == nb
collection_w.create_index(ct.default_float_vec_field_name, default_index_params, index_name="index_query")
collection_w.set_properties({'mmap.enabled': True})
collection_w.alter_index("index_query", {'mmap.enabled': True})
collection_w.load()
output_fields = [default_int_field_name, default_float_field_name, default_string_field_name]
expr = "varchar == \"\""
res, _ = collection_w.query(expr, output_fields=output_fields)
assert len(res) == nb
@pytest.mark.tags(CaseLabel.L1)
@pytest.mark.parametrize("expression", cf.gen_normal_string_expressions([default_string_field_name]))
def test_mmap_query_string_is_primary(self, expression):
"""
target: test query with output field only primary field
method: specify string primary field as output field
expected: return string primary field
"""
collection_w, vectors = self.init_collection_general(prefix, insert_data=True, is_index=False,
primary_field=ct.default_string_field_name)[0:2]
collection_w.set_properties({'mmap.enabled': True})
collection_w.create_index(ct.default_float_vec_field_name, default_index_params, index_name="query_expr_index")
collection_w.load()
res, _ = collection_w.query(expression, output_fields=[ct.default_string_field_name])
assert res[0].keys() == {ct.default_string_field_name}
@pytest.mark.tags(CaseLabel.L1)
def test_mmap_query_string_expr_with_prefixes(self):
"""
target: test query with prefix string expression
method: specify string is primary field, use prefix string expr
expected: verify query successfully
"""
collection_w, vectors = self.init_collection_general(prefix, insert_data=True,is_index=False,
primary_field=ct.default_string_field_name)[0:2]
collection_w.create_index(ct.default_float_vec_field_name, default_index_params, index_name="query_expr_pre_index")
collection_w.set_properties({'mmap.enabled': True})
collection_w.alter_index("query_expr_pre_index", {'mmap.enabled': True})
collection_w.load()
res = vectors[0].iloc[:1, :3].to_dict('records')
expression = 'varchar like "0%"'
output_fields = [default_int_field_name, default_float_field_name, default_string_field_name]
collection_w.query(expression, output_fields=output_fields,
check_task=CheckTasks.check_query_results, check_items={exp_res: res})
class TestQueryOperation(TestcaseBase):
"""

4
tests/python_client/testcases/test_search.py
View File

@ -5166,7 +5166,9 @@ class TestSearchBase(TestcaseBase):
output_fields = [default_string_field_name]
collection_w.search(binary_vectors[:default_nq], "binary_vector", search_params,
default_limit, default_search_string_exp, output_fields=output_fields,
check_task=CheckTasks.check_search_results)
check_task=CheckTasks.check_search_results,
check_items={"nq": nq,
"limit": ct.default_top_k})
class TestSearchDSL(TestcaseBase):