Support bulkload data in some popular format

[ShiKaiWi]

Describe This Problem

Currently, it is difficult to import data into ceresdb if the raw data is available in some popular format (CSV, parquet). And the only way is to parse the original data and write it to ceresdb by api.

Proposal

A bulkload api should be provided to do such things. User just tells ceresdb the data file path and its format, and ceresdb will do the bulkload work by itself.

Additional Context

Here is the similar feature provided by DuckDB: https://duckdb.org/docs/data/overview

[jiacai2050]

Previously discussed at https://github.com/CeresDB/ceresdb/issues/153#issuecomment-1207363994

[ShiKaiWi]

#925 is part of this issue.