[dataset] introduce mechanism to include(or even allow serialization and share) datasets in CSV and YAML which calls the importer embedded in backend

[wey-gu]

as talked with @hetao92 to enable a larger dataset for starter/sharing, a format of tgz/zip package(CSV data and importer yaml) could be introduced and parsed by the studio.

Furthermore, there then could be allowed to define(include) template queries, and markdown docs to be rendered so that users could share, and create datasets that help others to learn and work.

cc @slimshadylol0303 @hetao92