admin/milvus
1
0
Fork 0
mirror of https://gitee.com/milvus-io/milvus.git synced 2026-01-07 19:31:51 +08:00
Code Issues Packages Projects Releases Wiki Activity

[skip ci]Enable customize segment size test (#12092)

Browse Source 
Signed-off-by: yanliang567 <yanliang.qiao@zilliz.com>
This commit is contained in:
yanliang567 2021-11-19 10:21:13 +08:00 committed by GitHub
parent 2d81453ee0
commit 95eae52b19
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 88 additions and 83 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

171
tests/python_client/customize/test_customize_segment_size.py
View File

@ -4,6 +4,7 @@ import time
from pymilvus import connections
from utils.util_log import test_log as log
from base.collection_wrapper import ApiCollectionWrapper
from base.utility_wrapper import ApiUtilityWrapper
from common import common_func as cf
from common import common_type as ct
from milvus_operator import MilvusOperator
@ -11,7 +12,6 @@ from common.milvus_sys import MilvusSys
from common.common_type import CaseLabel
customize_segment_sizes = [128, 1024]
namespace = 'chaos-testing'
@ -38,25 +38,23 @@ def _install_milvus(seg_size):
return release_name, None, None
@pytest.mark.skip(reason="issue #11952")
class TestCustomizeSegmentSize:
def teardown_method(self):
pass
milvus_op = MilvusOperator()
milvus_op.uninstall(self.release_name, namespace)
connections.disconnect("default")
connections.remove_connection("default")
@pytest.mark.tags(CaseLabel.L3)
@pytest.mark.parametrize('size_id', [i for i in range(len(customize_segment_sizes))])
def test_customize_segment_size(self, size_id):
@pytest.mark.parametrize('seg_size, seg_count', [(128, 10), (1024, 2)])
def test_customize_segment_size(self, seg_size, seg_count):
"""
steps
1. [test_milvus_install]: set up milvus with customized simd configured
2. [test_simd_compat_e2e]: verify milvus is working well
4. [test_milvus_cleanup]: delete milvus instances in teardown
"""
size = customize_segment_sizes[size_id]
log.info(f"start to install milvus with segment size {size}")
release_name, host, port = _install_milvus(size)
log.info(f"start to install milvus with segment size {seg_size}")
release_name, host, port = _install_milvus(seg_size)
self.release_name = release_name
assert host is not None
conn = connections.connect("default", host=host, port=port)
@ -64,76 +62,83 @@ class TestCustomizeSegmentSize:
mil = MilvusSys(alias="default")
log.info(f"milvus build version: {mil.build_version}")
# log.info(f"start to e2e verification: {size}")
# # create
# name = cf.gen_unique_str("compat")
# t0 = time.time()
# collection_w = ApiCollectionWrapper()
# collection_w.init_collection(name=name,
# schema=cf.gen_default_collection_schema(),
# timeout=40)
# tt = time.time() - t0
# assert collection_w.name == name
# entities = collection_w.num_entities
# log.info(f"assert create collection: {tt}, init_entities: {entities}")
#
# # insert
# data = cf.gen_default_list_data()
# t0 = time.time()
# _, res = collection_w.insert(data)
# tt = time.time() - t0
# log.info(f"assert insert: {tt}")
# assert res
#
# # flush
# t0 = time.time()
# assert collection_w.num_entities == len(data[0]) + entities
# tt = time.time() - t0
# entities = collection_w.num_entities
# log.info(f"assert flush: {tt}, entities: {entities}")
#
# # search
# collection_w.load()
# search_vectors = cf.gen_vectors(1, ct.default_dim)
# search_params = {"metric_type": "L2", "params": {"nprobe": 16}}
# t0 = time.time()
# res_1, _ = collection_w.search(data=search_vectors,
# anns_field=ct.default_float_vec_field_name,
# param=search_params, limit=1)
# tt = time.time() - t0
# log.info(f"assert search: {tt}")
# assert len(res_1) == 1
# collection_w.release()
#
# # index
# d = cf.gen_default_list_data()
# collection_w.insert(d)
# log.info(f"assert index entities: {collection_w.num_entities}")
# _index_params = {"index_type": "IVF_SQ8", "params": {"nlist": 64}, "metric_type": "L2"}
# t0 = time.time()
# index, _ = collection_w.create_index(field_name=ct.default_float_vec_field_name,
# index_params=_index_params,
# name=cf.gen_unique_str())
# tt = time.time() - t0
# log.info(f"assert index: {tt}")
# assert len(collection_w.indexes) == 1
#
# # search
# t0 = time.time()
# collection_w.load()
# tt = time.time() - t0
# log.info(f"assert load: {tt}")
# search_vectors = cf.gen_vectors(1, ct.default_dim)
# t0 = time.time()
# res_1, _ = collection_w.search(data=search_vectors,
# anns_field=ct.default_float_vec_field_name,
# param=search_params, limit=1)
# tt = time.time() - t0
# log.info(f"assert search: {tt}")
#
# # query
# term_expr = f'{ct.default_int64_field_name} in [1001,1201,4999,2999]'
# t0 = time.time()
# res, _ = collection_w.query(term_expr)
# tt = time.time() - t0
# log.info(f"assert query result {len(res)}: {tt}")
log.info(f"start to e2e verification: {seg_size}")
# create
name = cf.gen_unique_str("segsiz")
t0 = time.time()
collection_w = ApiCollectionWrapper()
collection_w.init_collection(name=name,
schema=cf.gen_default_collection_schema(),
timeout=40)
tt = time.time() - t0
assert collection_w.name == name
entities = collection_w.num_entities
log.info(f"assert create collection: {tt}, init_entities: {entities}")
# insert
nb = 50000
data = cf.gen_default_list_data(nb=nb)
t0 = time.time()
_, res = collection_w.insert(data)
tt = time.time() - t0
log.info(f"assert insert: {tt}")
assert res
# insert 2 million entities
rounds = 40
for _ in range(rounds-1):
_, res = collection_w.insert(data)
entities = collection_w.num_entities
assert entities == nb * rounds
# load
collection_w.load()
utility_wrap = ApiUtilityWrapper()
segs, _ = utility_wrap.get_query_segment_info(collection_w.name)
log.info(f"assert segments: {len(segs)}")
assert len(segs) == seg_count
# search
search_vectors = cf.gen_vectors(1, ct.default_dim)
search_params = {"metric_type": "L2", "params": {"nprobe": 16}}
t0 = time.time()
res_1, _ = collection_w.search(data=search_vectors,
anns_field=ct.default_float_vec_field_name,
param=search_params, limit=1, timeout=30)
tt = time.time() - t0
log.info(f"assert search: {tt}")
assert len(res_1) == 1
collection_w.release()
# index
d = cf.gen_default_list_data()
collection_w.insert(d)
log.info(f"assert index entities: {collection_w.num_entities}")
_index_params = {"index_type": "IVF_SQ8", "params": {"nlist": 64}, "metric_type": "L2"}
t0 = time.time()
index, _ = collection_w.create_index(field_name=ct.default_float_vec_field_name,
index_params=_index_params,
name=cf.gen_unique_str(), timeout=120)
tt = time.time() - t0
log.info(f"assert index: {tt}")
assert len(collection_w.indexes) == 1
# search
t0 = time.time()
collection_w.load()
tt = time.time() - t0
log.info(f"assert load: {tt}")
search_vectors = cf.gen_vectors(1, ct.default_dim)
t0 = time.time()
res_1, _ = collection_w.search(data=search_vectors,
anns_field=ct.default_float_vec_field_name,
param=search_params, limit=1, timeout=30)
tt = time.time() - t0
log.info(f"assert search: {tt}")
# query
term_expr = f'{ct.default_int64_field_name} in [1001,1201,4999,2999]'
t0 = time.time()
res, _ = collection_w.query(term_expr, timeout=30)
tt = time.time() - t0
log.info(f"assert query result {len(res)}: {tt}")