2019-04-21 09:01:02 +00:00
|
|
|
////////////////////////////////////////////////////////////////////////////////
|
|
|
|
// Copyright 上海赜睿信息科技有限公司(Zilliz) - All Rights Reserved
|
|
|
|
// Unauthorized copying of this file, via any medium is strictly prohibited.
|
|
|
|
// Proprietary and confidential.
|
|
|
|
////////////////////////////////////////////////////////////////////////////////
|
|
|
|
#include <gtest/gtest.h>
|
2019-04-23 03:57:42 +00:00
|
|
|
#include <faiss/IndexFlat.h>
|
|
|
|
#include <faiss/MetaIndexes.h>
|
|
|
|
#include <faiss/AutoTune.h>
|
2019-04-23 12:55:15 +00:00
|
|
|
#include <thread>
|
|
|
|
#include <easylogging++.h>
|
2019-04-28 02:29:14 +00:00
|
|
|
#include <chrono>
|
2019-04-21 09:01:02 +00:00
|
|
|
|
|
|
|
#include "db/DB.h"
|
2019-04-23 03:57:42 +00:00
|
|
|
#include "faiss/Index.h"
|
2019-04-21 09:01:02 +00:00
|
|
|
|
|
|
|
using namespace zilliz::vecwise;
|
|
|
|
|
2019-04-28 02:29:14 +00:00
|
|
|
#define TIMING
|
|
|
|
|
|
|
|
#ifdef TIMING
|
|
|
|
#define INIT_TIMER auto start = std::chrono::high_resolution_clock::now();
|
|
|
|
#define START_TIMER start = std::chrono::high_resolution_clock::now();
|
|
|
|
#define STOP_TIMER(name) LOG(DEBUG) << "RUNTIME of " << name << ": " << \
|
|
|
|
std::chrono::duration_cast<std::chrono::milliseconds>( \
|
|
|
|
std::chrono::high_resolution_clock::now()-start \
|
|
|
|
).count() << " ms ";
|
|
|
|
#else
|
|
|
|
#define INIT_TIMER
|
|
|
|
#define START_TIMER
|
|
|
|
#define STOP_TIMER(name)
|
|
|
|
#endif
|
|
|
|
|
2019-04-26 09:59:51 +00:00
|
|
|
class DBTest : public ::testing::Test {
|
|
|
|
protected:
|
|
|
|
virtual void SetUp() {
|
|
|
|
el::Configurations defaultConf;
|
|
|
|
defaultConf.setToDefault();
|
|
|
|
defaultConf.set(el::Level::Debug,
|
|
|
|
el::ConfigurationType::Format, "[%thread-%datetime-%level]: %msg (%fbase:%line)");
|
|
|
|
el::Loggers::reconfigureLogger("default", defaultConf);
|
|
|
|
}
|
|
|
|
|
|
|
|
};
|
|
|
|
|
2019-04-21 09:01:02 +00:00
|
|
|
namespace {
|
|
|
|
void ASSERT_STATS(engine::Status& stat) {
|
|
|
|
ASSERT_TRUE(stat.ok());
|
|
|
|
if(!stat.ok()) {
|
|
|
|
std::cout << stat.ToString() << std::endl;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-04-26 09:59:51 +00:00
|
|
|
TEST_F(DBTest, DB_TEST) {
|
|
|
|
|
2019-04-21 09:01:02 +00:00
|
|
|
static const std::string group_name = "test_group";
|
|
|
|
static const int group_dim = 256;
|
|
|
|
|
|
|
|
engine::Options opt;
|
2019-04-23 12:55:15 +00:00
|
|
|
opt.memory_sync_interval = 1;
|
2019-04-28 05:56:29 +00:00
|
|
|
opt.index_trigger_size = 1024*group_dim;
|
2019-04-21 09:01:02 +00:00
|
|
|
opt.meta.backend_uri = "http://127.0.0.1";
|
2019-04-23 03:21:14 +00:00
|
|
|
opt.meta.path = "/tmp/vecwise_test/db_test";
|
2019-04-21 09:01:02 +00:00
|
|
|
|
|
|
|
engine::DB* db = nullptr;
|
|
|
|
engine::DB::Open(opt, &db);
|
|
|
|
ASSERT_TRUE(db != nullptr);
|
|
|
|
|
|
|
|
engine::meta::GroupSchema group_info;
|
|
|
|
group_info.dimension = group_dim;
|
|
|
|
group_info.group_id = group_name;
|
|
|
|
engine::Status stat = db->add_group(group_info);
|
|
|
|
|
|
|
|
engine::meta::GroupSchema group_info_get;
|
|
|
|
group_info_get.group_id = group_name;
|
|
|
|
stat = db->get_group(group_info_get);
|
|
|
|
ASSERT_STATS(stat);
|
|
|
|
ASSERT_EQ(group_info_get.dimension, group_dim);
|
|
|
|
|
|
|
|
engine::IDNumbers vector_ids;
|
2019-04-23 12:55:15 +00:00
|
|
|
engine::IDNumbers target_ids;
|
|
|
|
|
|
|
|
int d = 256;
|
2019-04-28 05:56:29 +00:00
|
|
|
int nb = 100;
|
2019-04-23 12:55:15 +00:00
|
|
|
float *xb = new float[d * nb];
|
|
|
|
for(int i = 0; i < nb; i++) {
|
|
|
|
for(int j = 0; j < d; j++) xb[d * i + j] = drand48();
|
|
|
|
xb[d * i] += i / 2000.;
|
2019-04-21 09:01:02 +00:00
|
|
|
}
|
|
|
|
|
2019-04-23 12:55:15 +00:00
|
|
|
int qb = 1;
|
|
|
|
float *qxb = new float[d * qb];
|
|
|
|
for(int i = 0; i < qb; i++) {
|
|
|
|
for(int j = 0; j < d; j++) qxb[d * i + j] = drand48();
|
|
|
|
qxb[d * i] += i / 2000.;
|
|
|
|
}
|
|
|
|
|
2019-04-28 05:56:29 +00:00
|
|
|
std::thread search([&]() {
|
|
|
|
engine::QueryResults results;
|
|
|
|
int k = 10;
|
|
|
|
std::this_thread::sleep_for(std::chrono::seconds(2));
|
|
|
|
/* std::this_thread::sleep_for(std::chrono::milliseconds(30)); */
|
|
|
|
|
|
|
|
INIT_TIMER;
|
|
|
|
std::stringstream ss;
|
|
|
|
long count = 0;
|
|
|
|
|
|
|
|
for (auto j=0; j<5; ++j) {
|
|
|
|
ss.str("");
|
|
|
|
db->count(group_name, count);
|
|
|
|
|
|
|
|
ss << "Search " << j << " With Size " << count;
|
|
|
|
|
|
|
|
START_TIMER;
|
|
|
|
stat = db->search(group_name, k, qb, qxb, results);
|
|
|
|
STOP_TIMER(ss.str());
|
|
|
|
|
|
|
|
ASSERT_STATS(stat);
|
|
|
|
ASSERT_EQ(results[0][0], target_ids[0]);
|
|
|
|
std::this_thread::sleep_for(std::chrono::seconds(1));
|
|
|
|
}
|
|
|
|
});
|
|
|
|
|
|
|
|
int loop = 40000;
|
2019-04-23 12:55:15 +00:00
|
|
|
|
|
|
|
for (auto i=0; i<loop; ++i) {
|
|
|
|
if (i==40) {
|
|
|
|
db->add_vectors(group_name, qb, qxb, target_ids);
|
|
|
|
} else {
|
|
|
|
db->add_vectors(group_name, nb, xb, vector_ids);
|
|
|
|
}
|
2019-04-28 05:56:29 +00:00
|
|
|
std::this_thread::sleep_for(std::chrono::microseconds(100));
|
2019-04-23 12:55:15 +00:00
|
|
|
}
|
2019-04-23 03:57:42 +00:00
|
|
|
|
2019-04-28 05:56:29 +00:00
|
|
|
search.join();
|
2019-04-23 12:55:15 +00:00
|
|
|
|
|
|
|
delete [] xb;
|
|
|
|
delete [] qxb;
|
|
|
|
delete db;
|
|
|
|
engine::DB::Open(opt, &db);
|
|
|
|
db->drop_all();
|
2019-04-23 03:57:42 +00:00
|
|
|
delete db;
|
2019-04-26 09:59:51 +00:00
|
|
|
};
|
2019-04-23 03:57:42 +00:00
|
|
|
|
2019-04-26 09:59:51 +00:00
|
|
|
TEST_F(DBTest, SEARCH_TEST) {
|
2019-04-23 03:57:42 +00:00
|
|
|
static const std::string group_name = "test_group";
|
|
|
|
static const int group_dim = 256;
|
|
|
|
|
|
|
|
engine::Options opt;
|
|
|
|
opt.meta.backend_uri = "http://127.0.0.1";
|
|
|
|
opt.meta.path = "/tmp/search_test";
|
|
|
|
opt.index_trigger_size = 100000 * group_dim;
|
|
|
|
opt.memory_sync_interval = 1;
|
|
|
|
opt.merge_trigger_number = 1;
|
|
|
|
|
|
|
|
engine::DB* db = nullptr;
|
|
|
|
engine::DB::Open(opt, &db);
|
|
|
|
ASSERT_TRUE(db != nullptr);
|
|
|
|
|
|
|
|
engine::meta::GroupSchema group_info;
|
|
|
|
group_info.dimension = group_dim;
|
|
|
|
group_info.group_id = group_name;
|
|
|
|
engine::Status stat = db->add_group(group_info);
|
|
|
|
//ASSERT_STATS(stat);
|
|
|
|
|
|
|
|
engine::meta::GroupSchema group_info_get;
|
|
|
|
group_info_get.group_id = group_name;
|
|
|
|
stat = db->get_group(group_info_get);
|
|
|
|
ASSERT_STATS(stat);
|
|
|
|
ASSERT_EQ(group_info_get.dimension, group_dim);
|
|
|
|
|
|
|
|
|
|
|
|
// prepare raw data
|
2019-04-24 08:57:15 +00:00
|
|
|
size_t nb = 250000;
|
2019-04-23 03:57:42 +00:00
|
|
|
size_t nq = 10;
|
|
|
|
size_t k = 5;
|
|
|
|
std::vector<float> xb(nb*group_dim);
|
|
|
|
std::vector<float> xq(nq*group_dim);
|
|
|
|
std::vector<long> ids(nb);
|
|
|
|
|
|
|
|
std::random_device rd;
|
|
|
|
std::mt19937 gen(rd());
|
|
|
|
std::uniform_real_distribution<> dis_xt(-1.0, 1.0);
|
|
|
|
for (size_t i = 0; i < nb*group_dim; i++) {
|
|
|
|
xb[i] = dis_xt(gen);
|
|
|
|
if (i < nb){
|
|
|
|
ids[i] = i;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
for (size_t i = 0; i < nq*group_dim; i++) {
|
|
|
|
xq[i] = dis_xt(gen);
|
|
|
|
}
|
|
|
|
|
|
|
|
// result data
|
|
|
|
//std::vector<long> nns_gt(k*nq);
|
|
|
|
std::vector<long> nns(k*nq); // nns = nearst neg search
|
|
|
|
//std::vector<float> dis_gt(k*nq);
|
|
|
|
std::vector<float> dis(k*nq);
|
|
|
|
|
|
|
|
// prepare ground-truth
|
|
|
|
//faiss::Index* index_gt(faiss::index_factory(group_dim, "IDMap,Flat"));
|
|
|
|
//index_gt->add_with_ids(nb, xb.data(), ids.data());
|
|
|
|
//index_gt->search(nq, xq.data(), 1, dis_gt.data(), nns_gt.data());
|
|
|
|
|
|
|
|
// insert data
|
|
|
|
const int batch_size = 100;
|
|
|
|
for (int j = 0; j < nb / batch_size; ++j) {
|
|
|
|
stat = db->add_vectors(group_name, batch_size, xb.data()+batch_size*j*group_dim, ids);
|
2019-04-24 08:57:15 +00:00
|
|
|
if (j == 200){ sleep(1);}
|
2019-04-23 03:57:42 +00:00
|
|
|
ASSERT_STATS(stat);
|
|
|
|
}
|
|
|
|
|
2019-04-24 08:57:15 +00:00
|
|
|
sleep(3); // wait until build index finish
|
2019-04-23 03:57:42 +00:00
|
|
|
|
2019-04-21 09:01:02 +00:00
|
|
|
engine::QueryResults results;
|
2019-04-23 03:57:42 +00:00
|
|
|
stat = db->search(group_name, k, nq, xq.data(), results);
|
2019-04-21 09:01:02 +00:00
|
|
|
ASSERT_STATS(stat);
|
2019-04-23 03:57:42 +00:00
|
|
|
|
|
|
|
// TODO(linxj): add groundTruth assert
|
2019-04-21 09:01:02 +00:00
|
|
|
|
|
|
|
delete db;
|
2019-04-24 08:57:15 +00:00
|
|
|
|
2019-04-23 12:55:15 +00:00
|
|
|
engine::DB::Open(opt, &db);
|
|
|
|
db->drop_all();
|
|
|
|
delete db;
|
2019-04-26 09:59:51 +00:00
|
|
|
};
|