def do_count(table_name, milvus_cli, mysql_cli):
if not table_name:
table_name = DEFAULT_TABLE
try:
if not milvus_cli.has_collection(table_name):
return None
milvus_num = milvus_cli.count(table_name)
mysql_num = mysql_cli.count_table(table_name)
LOGGER.debug("The num of Milvus: {} and Mysql: {}".format(
milvus_num, mysql_num))
assert milvus_num == mysql_num
return milvus_num
except Exception as e:
LOGGER.error(" Error with count table {}".format(e))
sys.exit(1)
def search_by_milvus_ids(self, ids, table_name):
self.test_connection()
str_ids = str(ids).replace('[', '').replace(']', '')
sql = "select * from " + table_name + " where milvus_id in (" + str_ids + ") order by field (milvus_id," + str_ids + ");"
try:
self.cursor.execute(sql)
results = self.cursor.fetchall()
results_id = [res[0] for res in results]
results_title = [res[1] for res in results]
results_text = [res[2] for res in results]
LOGGER.debug("MYSQL search by milvus id.")
return results_id, results_title, results_text
except Exception as e:
LOGGER.error("MYSQL ERROR: {} with sql: {}".format(e, sql))
sys.exit(1)
def insert(self, collection_name, vectors):
try:
self.create_collection(collection_name)
self.collection = Collection(name=collection_name)
data = [vectors]
mr = self.collection.insert(data)
ids = mr.primary_keys
self.collection.load()
LOGGER.debug(
"Insert vectors to Milvus in collection: {} with {} rows".
format(collection_name, len(vectors)))
return ids
except Exception as e:
LOGGER.error("Failed to load data to Milvus: {}".format(e))
sys.exit(1)
def do_drop(table_name, milvus_cli, mysql_cli):
if not table_name:
table_name = DEFAULT_TABLE
try:
if not milvus_cli.has_collection(table_name):
msg = "Milvus doesn't have a collection named {}".format(
table_name)
return msg
#return {'status': True, 'msg': msg}
status = milvus_cli.delete_collection(table_name)
mysql_cli.delete_table(table_name)
return status
except Exception as e:
LOGGER.error(" Error with drop table: {}".format(e))
sys.exit(1)
def create_index(self, collection_name):
try:
index_param = {'nlist': 16384}
status = self.client.create_index(collection_name,
IndexType.IVF_FLAT, index_param)
if not status.code:
LOGGER.debug(
"Successfully create index in collection:{} with param:{}".
format(collection_name, index_param))
return status
else:
raise Exception(status.message)
except Exception as e:
LOGGER.error("Failed to create index: {}".format(e))
sys.exit(1)
def insert(self, collection_name, vectors):
try:
self.create_colllection(collection_name)
status, ids = self.client.insert(collection_name=collection_name,
records=vectors)
if not status.code:
LOGGER.debug(
"Insert vectors to Milvus in collection: {} with {} rows".
format(collection_name, len(vectors)))
return ids
else:
raise Exception(status.message)
except Exception as e:
LOGGER.error("Failed to load data to Milvus: {}".format(e))
sys.exit(1)
def create_index(self, collection_name, index_params):
try:
self.set_collection(collection_name)
status = self.collection.create_index(field_name="embedding",
index_params=index_params)
if not status.code:
self.collection.load()
LOGGER.debug(
"Successfully create index in collection:{} with param:{}".
format(collection_name, index_params))
return status
else:
raise Exception(status.message)
except Exception as e:
LOGGER.error("Failed to create index: {}".format(e))
sys.exit(1)
def search_in_milvus(table_name, query_sentence,milvus_cli, mysql_cli):
if not table_name:
table_name = DEFAULT_TABLE
try:
query_data = [query_sentence]
vectors = bc.encode(query_data)
query_list = normaliz_vec(vectors.tolist())
LOGGER.info("Successfully insert query list")
results = milvus_cli.search_vectors(table_name,query_list,TOP_K)
vids = [str(x.id) for x in results[0]]
print("-----------------", vids)
ids,title,text= mysql_cli.search_by_milvus_ids(vids, table_name)
distances = [x.distance for x in results[0]]
return ids,title, text, distances
except Exception as e:
LOGGER.error(" Error with search : {}".format(e))
sys.exit(1)
def create_colllection(self, collection_name):
try:
if not self.has_collection(collection_name):
collection_param = {
'collection_name': collection_name,
'dimension': VECTOR_DIMENSION,
'index_file_size': INDEX_FILE_SIZE,
'metric_type': METRIC_TYPE
}
status = self.client.create_collection(collection_param)
if status.code != 0:
raise Exception(status.message)
LOGGER.debug(
"Create Milvus collection: {}".format(collection_name))
except Exception as e:
LOGGER.error("Failed to load data to Milvus: {}".format(e))
sys.exit(1)
def do_search(host, table_name, img_path, model, milvus_client, mysql_cli):
if not table_name:
table_name = DEFAULT_TABLE
try:
feat = model.resnet50_extract_feat(img_path)
vectors = milvus_client.search_vectors(table_name, [feat], TOP_K)
vids = [str(x.id) for x in vectors[0]]
paths = mysql_cli.search_by_milvus_ids(vids, table_name)
distances = [x.distance for x in vectors[0]]
for i in range(len(paths)):
tmp = "http://" + str(host) + "/data?gif_path=" + str(paths[i])
paths[i] = tmp
return paths, distances
except Exception as e:
LOGGER.error(" Error with search : {}".format(e))
sys.exit(1)
def search_vectors(self, collection_name, vectors, top_k):
try:
search_param = {'nprobe': 16}
status, result = self.client.search(
collection_name=collection_name,
query_records=vectors,
top_k=top_k,
params=search_param)
if not status.code:
LOGGER.debug("Successfully search in collection: {}".format(
collection_name))
return result
else:
raise Exception(status.message)
except Exception as e:
LOGGER.error("Failed to search vectors in Milvus: {}".format(e))
sys.exit(1)
async def search_images(image: UploadFile = File(...), table_name: str = None):
# Search the upload image in Milvus/MySQL
try:
# Save the upload image to server.
content = await image.read()
print('read pic succ')
img_path = os.path.join(UPLOAD_PATH, image.filename)
with open(img_path, "wb+") as f:
f.write(content)
paths, distances = do_search(table_name, img_path, MODEL, MILVUS_CLI,
MYSQL_CLI)
res = dict(zip(paths, distances))
res = sorted(res.items(), key=lambda item: item[1])
LOGGER.info("Successfully searched similar images!")
return res
except Exception as e:
LOGGER.error(e)
return {'status': False, 'msg': e}, 400
def extract_features(img_dir, model):
try:
cache = Cache('./tmp')
feats = []
names = []
img_list = get_imgs(img_dir)
total = len(img_list)
cache['total'] = total
for i, img_path in enumerate(img_list):
norm_feat = model.resnet50_extract_feat(img_path)
feats.append(norm_feat)
names.append(img_path.encode())
cache['current'] = i + 1
print("Extracting feature from image No. %d , %d images in total" %
(i + 1, total))
return feats, names
except Exception as e:
LOGGER.error(" Error with extracting feature from image {}".format(e))
sys.exit(1)
def search_vectors(self, collection_name, vectors, top_k):
try:
self.set_collection(collection_name)
search_params = {
"metric_type": METRIC_TYPE,
"params": {
"nprobe": 16
}
}
# data = [vectors]
res = self.collection.search(vectors,
anns_field="embedding",
param=search_params,
limit=top_k)
LOGGER.debug("Successfully search in collection: {}".format(res))
return res
except Exception as e:
LOGGER.error("Failed to search vectors in Milvus: {}".format(e))
sys.exit(1)
async def do_search_api(table_name: str = None, query_sentence: str = None):
try:
ids, results_classes, seq_genes, distances = search_in_milvus(
table_name, query_sentence, MILVUS_CLI, MYSQL_CLI)
res = []
for i, c, s, d in zip(ids, results_classes, seq_genes, distances):
dicts = {
'milvus_id': i,
'seq_class': c,
'seq_gene': s,
'IP distance': d
}
res += [dicts]
LOGGER.info("Successfully searched similar sequence!")
return res
except Exception as e:
LOGGER.error(e)
return {'status': False, 'msg': e}, 400
async def load_text(file: UploadFile = File(...), table_name: str = None):
try:
text = await file.read()
fname = file.filename
dirs = "data"
if not os.path.exists(dirs):
os.makedirs(dirs)
fname_path = os.path.join(os.getcwd(), os.path.join(dirs, fname))
with open(fname_path, 'wb') as f:
f.write(text)
except Exception as e:
return {'status': False, 'msg': 'Failed to load data.'}
# Insert all the image under the file path to Milvus/MySQL
try:
total_num = import_data(table_name, fname_path, MILVUS_CLI, MYSQL_CLI)
LOGGER.info(
"Successfully loaded data, total count: {}".format(total_num))
return "Successfully loaded data!"
except Exception as e:
LOGGER.error(e)
return {'status': False, 'msg': e}, 400
def search_in_milvus(table_name, query_sentence, milvus_cli, mysql_cli):
if not table_name:
table_name = DEFAULT_TABLE
try:
kmers = build_kmers(query_sentence,KMER_K)
query_data = [" ".join(kmers)]
query_list = encode_seq(query_data)
LOGGER.info("Searching...")
results = milvus_cli.search_vectors(table_name,query_list,TOP_K)
vids = [str(x.id) for x in results[0]]
print("-----------------", vids)
ids, results_classes = mysql_cli.search_by_milvus_ids(vids, table_name)
distances = [x.distance for x in results[0]]
df_class = pd.read_table(SEQ_CLASS_PATH)
class_dict = dict()
for i in range(len(df_class)):
class_dict[df_class['class'][i]] = df_class['gene_family'][i]
seq_genes = [class_dict[int(x)] for x in results_classes]
return ids, results_classes, seq_genes, distances
except Exception as e:
LOGGER.error(" Error with search : {}".format(e))
sys.exit(1)
def extract_features(video_dir, model, frame):
try:
cache = Cache('./tmp')
feats = []
names = []
video_list = get_video(video_dir)
total = len(video_list)
cache['total'] = total
for i, video_path in enumerate(video_list):
imgs = frame.extract_frame(video_path)
for img_path in imgs:
norm_feat = model.resnet50_extract_feat(img_path)
feats.append(norm_feat)
names.append(video_path.encode())
cache['current'] = i + 1
print(
"%d video in total, extracting feature from video No. %d , and the video has %d frames."
% (i + 1, total, len(imgs)))
return feats, names
except Exception as e:
LOGGER.error(" Error with extracting feature from image {}".format(e))
sys.exit(1)
def create_index(self, collection_name):
try:
self.set_collection(collection_name)
default_index = {
"index_type": "IVF_SQ8",
"metric_type": METRIC_TYPE,
"params": {
"nlist": 16384
}
}
status = self.collection.create_index(field_name="embedding",
index_params=default_index)
if not status.code:
LOGGER.debug(
"Successfully create index in collection:{} with param:{}".
format(collection_name, default_index))
return status
else:
raise Exception(status.message)
except Exception as e:
LOGGER.error("Failed to create index: {}".format(e))
sys.exit(1)
async def search_images(request: Request,
image: UploadFile = File(...),
table_name: str = None):
# Search the upload image in Milvus/MySQL
try:
# Save the upload image to server.
content = await image.read()
img_path = os.path.join(UPLOAD_PATH, image.filename)
with open(img_path, "wb+") as f:
f.write(content)
host = request.headers['host']
paths, distances = do_search(host, table_name, img_path, MODEL,
MILVUS_CLI, MYSQL_CLI)
res = {}
for p, d in zip(paths, distances):
if not p in res or res[p] > d:
res[p] = d
res = sorted(res.items(), key=lambda item: item[1])
LOGGER.info("Successfully searched similar images!")
return res
except Exception as e:
LOGGER.error(e)
return {'status': False, 'msg': e}, 400
def create_collection(self, collection_name):
try:
if not self.has_collection(collection_name):
field1 = FieldSchema(name="id",
dtype=DataType.INT64,
descrition="int64",
is_primary=True,
auto_id=True)
field2 = FieldSchema(name="embedding",
dtype=DataType.FLOAT_VECTOR,
descrition="float vector",
dim=VECTOR_DIMENSION,
is_primary=False)
schema = CollectionSchema(fields=[field1, field2],
description="collection description")
self.collection = Collection(name=collection_name,
schema=schema)
LOGGER.debug("Create Milvus collection: {}".format(
self.collection))
return "OK"
except Exception as e:
LOGGER.error("Failed to load data to Milvus: {}".format(e))
sys.exit(1)
async def do_load_api(file: UploadFile = File(...), table_name: str = None):
try:
text = await file.read()
fname = file.filename
dirs = "QA_data"
if not os.path.exists(dirs):
os.makedirs(dirs)
fname_path = os.path.join(os.getcwd(), os.path.join(dirs, fname))
with open(fname_path, 'wb') as f:
f.write(text)
except Exception as e:
return {'status': False, 'msg': 'Failed to load data.'}
try:
total_num = do_load(table_name, fname_path, MODEL, MILVUS_CLI,
MYSQL_CLI)
LOGGER.info(
"Successfully loaded data, total count: {}".format(total_num))
return {
'status': True,
'msg': "Successfully loaded data: {}".format(total_num)
}, 200
except Exception as e:
LOGGER.error(e)
return {'status': False, 'msg': e}, 400
def do_search(table_name, img_path, model, milvus_client, mysql_cli):
try:
if not table_name:
table_name = DEFAULT_TABLE
detector = Detector()
run(detector, img_path)
vecs = get_object_vector(model, img_path + '/object')
# feat = model.resnet50_extract_feat(img_path)
results = milvus_client.search_vectors(table_name, vecs, TOP_K)
ids = []
distances = []
for result in results:
for j in result:
ids.append(j.id)
distances.append(j.distance)
# res_id = [x for x in query_name_from_ids(vids)]
# vids = [str(x.id) for x in vectors[0]]
paths = mysql_cli.search_by_milvus_ids(ids, table_name)
# distances = [x.distance for x in vectors[0]]
shutil.rmtree(img_path)
return paths, distances
except Exception as e:
LOGGER.error(" Error with search : {}".format(e))
sys.exit(1)
def has_collection(self, collection_name):
try:
return utility.has_collection(collection_name)
except Exception as e:
LOGGER.error("Failed to load data to Milvus: {}".format(e))
sys.exit(1)
