milvus/pyengine/tests/test_function.py

103 lines
2.7 KiB
Python

import numpy as np
import requests
import pytest
import logging
import json
url = "http://127.0.0.1:5000"
class TestEngineFunction():
def test_1m_add(self):
d = 4
nb = 100
nq = 1
k = 10
_, xb, xq = get_dataset(d, nb, 1, nq)
groupid = "test_search_3"
route_group = url + "/vector/group/" + groupid
r = requests.post(route_group, json={"dimension": d})
# import dataset
vector_add_route = url + "/vector/add/" + groupid
for i in xb:
data = dict()
data['vector'] = i.tolist()
# print(data)
r = requests.post(vector_add_route, json=data)
print(r.json())
# search dataset
vector_search_route = url + "/vector/search/" + groupid
data = dict()
for i in xq:
data['vector'] = i.tolist()
data['limit'] = k
# print(data)
r = requests.get(vector_search_route, json=data)
print(r.json())
def test_restful_interface(self):
d = 4
nb = 100
nq = 1
k = 10
_, xb, xq = get_dataset(d, nb, 1, nq)
groupid_1 = "Group_1"
groupid_2 = "Group_2"
vector_add_route = url + "/vector/add/"
vector_search_route = url + "/vector/search/"
group_route = url + "/vector/group/"
group_list_route = url + "/vector/group"
# Add groupid
r = requests.post(group_route + groupid_1, json={"dimension": d})
print(r.json())
r = requests.post(group_route + groupid_2, json={"dimension": d})
print(r.json())
# Get groupid list
r = requests.get(group_list_route)
print(r.json())
# delete groupid
r = requests.delete(group_route + groupid_2)
print(r.json())
# get groupid
r = requests.get(group_route + groupid_1)
print(r.json())
# add vector
for i in xb:
data = dict()
data['vector'] = i.tolist()
# print(data)
r = requests.post(vector_add_route + groupid_1, json=data)
print(r.json())
# search dataset
data = dict()
for i in xq:
data['vector'] = i.tolist()
data['limit'] = k
# print(data)
r = requests.get(vector_search_route + groupid_1, json=data)
print(r.json())
def get_dataset(d, nb, nt, nq):
d1 = 10 # intrinsic dimension (more or less)
n = nb + nt + nq
rs = np.random.RandomState(1338)
x = rs.normal(size=(n, d1))
x = np.dot(x, rs.rand(d1, d))
x = x * (rs.rand(d) * 4 + 0.1)
x = np.sin(x)
x = x.astype('float32')
return x[:nt], x[nt:-nq], x[-nq:]