Improve pgpass handling

add scd by hash

Add sqlExec and inputBulkCsv methods

They allow low level sql and loading any csv into postgres

Hangle arrays

Change constructor

improve temporary tables

Add more parallelism for multiline csv

Add a splitFactor argument for inputBulk

Upgrade documentation and project version

Change the API to something like builder

- Optimize the parallelism
- make use of csv interface for both input/output bulk

Implement parallel "copy from".

This produces as many csv as the number of partition needed

Add hdfs handling

First release