setup clean base for Pietro

[vincent-octo]

This is a PR that sets up a clean python sub-project engine inside our existing codebase.

TODO

• Base structure that takes the Parquet file from the intake stage, get the columns from config, and outputs a Parquet file.

• Parallelize the processing step

• Put intake config into shared config

  I think it's better to have the shared config file (src/kanta/config.py) have only the variables that are meant for cross-module use.
  Other variables, e.g. chunking knobs, should belong to their own module.
  That way the shared config file doesn't turn into a gigantic mess.
  🤷

• logging

How it works

• The shared config is at the root of our package at src/kanta/config.py.
• The engine module is made available by having the file src/kanta/engine/__init__.py
• In src/kanta/engine/__main__.py the engine imports the injection module and the shared config with:

  from kanta import config
  from kanta.engine import injection
  
  if __name__ == '__main__':
      injection.fake_run(config.EXAMPLE_VAR)

• Running python3 -m kanta.engine will call src/kanta/engine/__main__.py

[vincent-octo]

@piotor87 Here is an example on how to have a clean setup.
Let me know if you have questions or want to discuss =)