Create CLI for csvqb

[robons]

You should be able to provide an info.json file along with a tidy-data CSV and the utility should generate the resulting CSV-W.

The aim of this work is to make it easier for Data Managers to iteratively change the info.json file and view the resulting CSV-W metadata without having to re-run the whole generation script. The CSV should effectively be an artifact of the previous build step.

Regarding validation errors: these should be output to the command line (with suitable colours and exit code for the CLI). There should also be a command line option to write the errors to a file named validation-errors.json (which should be cleanred & overwritten on each run).

[robons]

We'll eventually have this running as a command line task in the databaker-docker container once we've moved this functionality over into the gss-utils package.