def test_string_fresh():
DATA_STRS = [
"xyz", "beagcfa", "cea", "cb", "d", "c", "bdaf", "ddcd", "egbfa", "a",
"fba", "bcccfe", "ab", "bfgbfdc", "bcbbgf", "bfbb"
]
QUERY_STRS = ["abc", "def", "ghik"]
space_type = 'leven'
space_param = []
method_name = 'small_world_rand'
index_name = method_name + '.index'
index = nmslib_vector.init(space_type, space_param, method_name,
nmslib_vector.DataType.STRING,
nmslib_vector.DistType.INT)
for id, data in enumerate(DATA_STRS):
nmslib_vector.addDataPoint(index, id, data)
index_param = ['NN=17', 'initIndexAttempts=3', 'indexThreadQty=4']
query_time_param = ['initSearchAttempts=3']
nmslib_vector.createIndex(index, index_param)
nmslib_vector.setQueryTimeParams(index, query_time_param)
print 'Query time parameters are set'
print "Results for the freshly created index:"
k = 2
for idx, data in enumerate(QUERY_STRS):
print idx, nmslib_vector.knnQuery(index, k, data)
nmslib_vector.saveIndex(index, index_name)
print "The index %s is saved" % index_name
nmslib_vector.freeIndex(index)
def test_vector_loaded():
space_type = 'cosinesimil'
space_param = []
method_name = 'small_world_rand'
index_name = method_name + '.index'
index = nmslib_vector.init(
space_type,
space_param,
method_name,
nmslib_vector.DataType.VECTOR,
nmslib_vector.DistType.FLOAT)
for id, data in enumerate(read_data('sample_dataset.txt')):
nmslib_vector.addDataPoint(index, id, data)
query_time_param = ['initSearchAttempts=3']
nmslib_vector.loadIndex(index, index_name)
print "The index %s is loaded" % index_name
nmslib_vector.setQueryTimeParams(index,query_time_param)
print 'Query time parameters are set'
print "Results for the loaded index"
k = 2
for idx, data in enumerate(read_data('sample_queryset.txt')):
print idx, nmslib_vector.knnQuery(index, k, data)
nmslib_vector.freeIndex(index)
def test_vector_loaded():
space_type = 'cosinesimil'
space_param = []
method_name = 'small_world_rand'
index_name = method_name + '.index'
index = nmslib_vector.init(space_type, space_param, method_name,
nmslib_vector.DataType.VECTOR,
nmslib_vector.DistType.FLOAT)
for id, data in enumerate(read_data('sample_dataset.txt')):
nmslib_vector.addDataPoint(index, id, data)
query_time_param = ['initSearchAttempts=3']
nmslib_vector.loadIndex(index, index_name)
print "The index %s is loaded" % index_name
nmslib_vector.setQueryTimeParams(index, query_time_param)
print 'Query time parameters are set'
print "Results for the loaded index"
k = 2
for idx, data in enumerate(read_data('sample_queryset.txt')):
print idx, nmslib_vector.knnQuery(index, k, data)
nmslib_vector.freeIndex(index)
def fit(self, X):
os.environ['OMP_THREAD_LIMIT'] = '40'
import nmslib_vector
if self._method_name == 'vptree':
# To avoid this issue:
# terminate called after throwing an instance of 'std::runtime_error'
# what(): The data size is too small or the bucket size is too big. Select the parameters so that <total # of records> is NOT less than <bucket size> * 1000
# Aborted (core dumped)
self._index_param.append('bucketSize=%d' % min(int(X.shape[0] * 0.0005), 1000))
self._index = nmslib_vector.init(self._nmslib_metric, [], self._method_name, nmslib_vector.DataType.VECTOR, nmslib_vector.DistType.FLOAT)
for i, x in enumerate(X):
nmslib_vector.addDataPoint(self._index, i, x.tolist())
if os.path.exists(self._index_name):
print "Loading index from file"
nmslib_vector.loadIndex(self._index, self._index_name)
else:
nmslib_vector.createIndex(self._index, self._index_param)
# nmslib_vector.saveIndex(self._index, self._index_name)
nmslib_vector.setQueryTimeParams(self._index, self._query_param)
os.environ['OMP_THREAD_LIMIT'] = '1'
def load_index(col_name, category, index_version, reindex=False):
space_type = 'jsdivslow'
space_param = []
method_name = 'small_world_rand'
index_name = 'indexes/' + col_name + '_' + category + index_version + '.index'
file_exists = isfile(index_name)
if reindex or not file_exists:
build_n_save(col_name, category, index_version)
index = nmslib_vector.init(space_type, space_param, method_name,
nmslib_vector.DataType.VECTOR,
nmslib_vector.DistType.FLOAT)
nmslib_index = 'nmslib_index' + index_version
all_items_in_category = db[col_name].find({
'categories': category,
nmslib_index: {
'$exists': 1
}
})
t1 = time()
for idx, item in enumerate(all_items_in_category):
fp = item['fingerprint']
if type(fp) == list:
color = fp
elif type(fp) == dict:
color = fp['color']
else:
print('else')
continue
nmslib_vector.addDataPoint(index, idx, color)
# item_id = item['_id']
# db[col_name].update_one({'_id':item_id}, {'$set': {'nmslib_index': idx}})
t2 = time()
print('addDataPoints took %s secs' % str(t2 - t1))
# index_param = ['NN=17', 'initIndexAttempts=3', 'indexThreadQty=32']
query_time_param = ['initSearchAttempts=3']
nmslib_vector.loadIndex(index, index_name)
print "The index %s is loaded" % index_name
t3 = time()
print('createIndex took %s secs' % str(t3 - t2))
nmslib_vector.setQueryTimeParams(index, query_time_param)
return index, nmslib_vector
def create_index(col_name, category):
space_type = 'jsdivslow'
space_param = []
method_name = 'small_world_rand'
index_name = col_name + '_' + category + '.index'
index = nmslib_vector.init(space_type, space_param, method_name,
nmslib_vector.DataType.VECTOR,
nmslib_vector.DistType.FLOAT)
all_items_in_category = db[col_name].find({'categories': category})
t1 = time()
for idx, item in enumerate(all_items_in_category):
fp = item['fingerprint']
if type(fp) == list:
color = fp
elif type(fp) == dict:
color = fp['color']
else:
print('else')
continue
nmslib_vector.addDataPoint(index, idx, color)
item_id = item['_id']
db[col_name].update_one({'_id': item_id},
{'$set': {
'nmslib_index': idx
}})
t2 = time()
print('addDataPoints took %s secs' % str(t2 - t1))
index_param = ['NN=17', 'initIndexAttempts=3', 'indexThreadQty=32']
query_time_param = ['initSearchAttempts=3']
nmslib_vector.createIndex(index, index_param)
t3 = time()
print('createIndex took %s secs' % str(t3 - t2))
nmslib_vector.setQueryTimeParams(index, query_time_param)
nmslib_vector.saveIndex(index, index_name)
return index, nmslib_vector
def test_string_fresh():
DATA_STRS = ["xyz", "beagcfa", "cea", "cb",
"d", "c", "bdaf", "ddcd",
"egbfa", "a", "fba", "bcccfe",
"ab", "bfgbfdc", "bcbbgf", "bfbb"
]
QUERY_STRS = ["abc", "def", "ghik"]
space_type = 'leven'
space_param = []
method_name = 'small_world_rand'
index_name = method_name + '.index'
index = nmslib_vector.init(
space_type,
space_param,
method_name,
nmslib_vector.DataType.STRING,
nmslib_vector.DistType.INT)
for id, data in enumerate(DATA_STRS):
nmslib_vector.addDataPoint(index, id, data)
index_param = ['NN=17', 'initIndexAttempts=3', 'indexThreadQty=4']
query_time_param = ['initSearchAttempts=3']
nmslib_vector.createIndex(index, index_param)
nmslib_vector.setQueryTimeParams(index, query_time_param)
print 'Query time parameters are set'
print "Results for the freshly created index:"
k = 2
for idx, data in enumerate(QUERY_STRS):
print idx, nmslib_vector.knnQuery(index, k, data)
nmslib_vector.saveIndex(index, index_name)
print "The index %s is saved" % index_name
nmslib_vector.freeIndex(index)
def init_index(idx, k=10):
index_param = ['NN=%d' % k, 'initIndexAttempts=3', 'indexThreadQty=8']
query_time_param = ['initSearchAttempts=3']
nmslib_vector.createIndex(idx[0], index_param)
nmslib_vector.setQueryTimeParams(idx[0], query_time_param)
idx[2] = k