Jupyter工具库
通过json配置的schema, 轻松实现可视化参数配置, 可以触发事件交互, 可生成key-value格式,或者json格式.
# _IMPORT_('import github.com/qrsforever/nb_easy/easy_widget')
_IMPORT_('import gitee.com/qrsforever/nb_easy/easy_widget')
schema = {
'type': 'page',
'objs': [
{
'type': 'H',
'objs': [
nbeasy_widget_string('task', 'Task', 'zmq.repnet_tf.inference', width='30%', readonly=True),
nbeasy_widget_string('cfg.pigeon.msgkey', 'Mssage Key', 'zmq.repnet_tf.test', width='30%', readonly=True),
nbeasy_widget_string('__cfg.race_url', 'Race URL', 'http://host:port', width='30%', readonly=True),
]
},
nbeasy_widget_int('cfg.epoch', 'Epoch', 10, min_=10),
nbeasy_widget_float('cfg.lr', 'LearnRate', 0.001)
]
}
easy = nbeasy_schema_parse(schema, events=None, debug=True)
key-value输出(print(json.dumps(easy.get_all_kv(), indent=4))
:
{
"task": "zmq.repnet_tf.inference",
"cfg.pigeon.msgkey": "zmq.repnet_tf.test",
"cfg.epoch": 10,
"cfg.lr": 0.001
}
json-conf输出(print(json.dumps(easy.get_all_json(), indent=4))
:
{
"task": "zmq.repnet_tf.inference",
"cfg": {
"pigeon": {
"msgkey": "zmq.repnet_tf.test"
},
"epoch": 10,
"lr": 0.001
}
}