dataset configuration

Users who downloaded original dataset, can pull dsdl file and modify it's location configuration, it works too.

we use config.py to set dataset's location configuration, it supports two type locations:

1. load form local： set type = LocalFileReader, and change working_dir to actual path.
2. load from ali-oss： set type = AliOSSFileReader, then input your identification such as access_key_secret, endpoint, access_key_id, you also need to set bucket_name and working_dir so the loader can find your dataset.

an example of config.py file shown as:

local = dict(
    type="LocalFileReader",
    working_dir="local path of your media",
)

ali_oss = dict(
    type="AliOSSFileReader",
    access_key_secret="your secret key of aliyun oss",
    endpoint="your endpoint of aliyun oss",
    access_key_id="your access key of aliyun oss",
    bucket_name="your bucket name of aliyun oss",
    working_dir="the relative path of your media dir in the bucket")