def connect(url=None,
schema=None,
reflect_metadata=True,
engine_kwargs=None,
reflect_views=True,
ensure_schema=True,
row_type=row_type):
""" Opens a new connection to a database.
*url* can be any valid `SQLAlchemy engine URL`_. If *url* is not defined
it will try to use *DATABASE_URL* from environment variable. Returns an
instance of :py:class:`Database <dataset.Database>`. Set *reflect_metadata*
to False if you don't want the entire database schema to be pre-loaded.
This significantly speeds up connecting to large databases with lots of
tables. *reflect_views* can be set to False if you don't want views to be
loaded. Additionally, *engine_kwargs* will be directly passed to
SQLAlchemy, e.g. set *engine_kwargs={'pool_recycle': 3600}* will avoid `DB
connection timeout`_. Set *row_type* to an alternate dict-like class to
change the type of container rows are stored in.::
db = dataset.connect('sqlite:///factbook.db')
.. _SQLAlchemy Engine URL: http://docs.sqlalchemy.org/en/latest/core/engines.html#sqlalchemy.create_engine
.. _DB connection timeout: http://docs.sqlalchemy.org/en/latest/core/pooling.html#setting-pool-recycle
"""
if url is None:
url = os.environ.get('DATABASE_URL', 'sqlite://')
return Database(url,
schema=schema,
reflect_metadata=reflect_metadata,
engine_kwargs=engine_kwargs,
reflect_views=reflect_views,
ensure_schema=ensure_schema,
row_type=row_type)
def connect(
url=None,
schema=None,
engine_kwargs=None,
ensure_schema=True,
row_type=row_type,
sqlite_wal_mode=True,
on_connect_statements=None,
):
"""Opens a new connection to a database.
*url* can be any valid `SQLAlchemy engine URL`_. If *url* is not defined
it will try to use *DATABASE_URL* from environment variable. Returns an
instance of :py:class:`Database <dataset.Database>`. Additionally,
*engine_kwargs* will be directly passed to SQLAlchemy, e.g. set
*engine_kwargs={'pool_recycle': 3600}* will avoid `DB connection timeout`_.
Set *row_type* to an alternate dict-like class to change the type of
container rows are stored in.::
db = dataset.connect('sqlite:///factbook.db')
One of the main features of `dataset` is to automatically create tables and
columns as data is inserted. This behaviour can optionally be disabled via
the `ensure_schema` argument. It can also be overridden in a lot of the
data manipulation methods using the `ensure` flag.
If you want to run custom SQLite pragmas on database connect, you can add them
to on_connect_statements as a set of strings. You can view a full
`list of PRAGMAs here`_.
.. _SQLAlchemy Engine URL: http://docs.sqlalchemy.org/en/latest/core/engines.html#sqlalchemy.create_engine
.. _DB connection timeout: http://docs.sqlalchemy.org/en/latest/core/pooling.html#setting-pool-recycle
.. _list of PRAGMAs here: https://www.sqlite.org/pragma.html
"""
if url is None:
url = os.environ.get("DATABASE_URL", "sqlite://")
return Database(
url,
schema=schema,
engine_kwargs=engine_kwargs,
ensure_schema=ensure_schema,
row_type=row_type,
sqlite_wal_mode=sqlite_wal_mode,
on_connect_statements=on_connect_statements,
)
parser.add_argument('--cfg', type=str, default='config/default.yaml', help='*.cfg path')
opt = parser.parse_args()
with open(opt.cfg, encoding='utf-8') as f:
cfg = yaml.load(f, Loader=yaml.FullLoader) # model dict
os.chdir(cfg['root_dir'])
experiment_dir = os.path.join(cfg['root_dir'], cfg['experiment_name'])
if not os.path.exists(experiment_dir):
os.mkdir(experiment_dir)
train_data_csv = os.path.join(experiment_dir, 'ir_train.csv')
test_data_csv = os.path.join(experiment_dir, 'ir_test.csv')
Database(data_path=cfg['ir_train_dir'], save_path=train_data_csv)
Database(data_path=cfg['ir_test_dir'], save_path=test_data_csv)
save_m = os.path.join(experiment_dir, cfg['model_name']+'.h5') # 保存训练模型路径
save_h = os.path.join(experiment_dir, cfg['model_name']+'_train_history') # 保存训练历史路径
IMAGE_SIZE = cfg['image_size'] # 修改model中的尺寸
BATCH = cfg['batch_size']
EPOCH = cfg['epoch']
CLASS_MODE = cfg['class_mode']
COLOR_MODE = cfg['color_mode']
LR = cfg['lr']
# 模型
os.environ["CUDA_VISIBLE_DEVICES"] = str(cfg['gputouse'])
adam = Adam(lr=LR)
d_type = cfg['d_type'] # 距离类型
depth = cfg['depth'] # 检索返回深度, None为返回全部, P11 需要设置depth=None
feature_path = os.path.join(experiment_dir, 'features')
if not os.path.exists(feature_path):
os.mkdir(feature_path)
Q_PATH = cfg['q_path'] # 待检索数据集
S_PATH = cfg['s_path'] # 检索数据集
q_name = cfg['q_name']
s_name = cfg['s_name']
save_Q_csv = os.path.join(experiment_dir, 'ir_Q.csv')
save_S_csv = os.path.join(experiment_dir, 'ir_S.csv')
save_result_txt = os.path.join(experiment_dir, 'ir_query_result.txt')
qdb = Database(data_path=Q_PATH, save_path=save_Q_csv)
sdb = Database(data_path=S_PATH, save_path=save_S_csv)
model = load_irmodels(cfg)
feat_model = Model(inputs=model.input, outputs=model.get_layer(pick_layer).output)
print(feat_model.summary())
# 特征提取/加载
qsamlpe = feature_samples(model_name=model_name,
pick_layer=pick_layer,
model=feat_model,
db=qdb,
image_size=image_size,
color_mode=COLOR_MODE,
dataset=q_name,
