Issue 73 checkpoints

[galamit86]

This PR:

• Refactors the package, dividing utils into 4 submodules:
  • gcpl module contains the functions that interact with the GCP (named gcpl as gcp is used in the code already to indicate "google cloud project")
  • statline module contains functions that interact with Statline.
  • utils module contains generic functions (like create_dir, etc.)
  • main module contains the main logic of processing datasets.
• Per submodule, marks private functions, by adding an underscore. The public functions per module are:
  • gcpl:
  • upload_to_gcs
  • gcs_to_gbq
  • statline:
  • get_metadata_cbs
  • dataset_to_parquet (changed here from tables_to_parquet)
  • utils:
  • dict_to_json_file
  • convert_ndjsons_to_parquet
  • clean_python_name
  • main:
    • main
• Adds endpoint parameter, to allow for 3 options to end with when running main.main():
  • local - locally stored parquet files
  • gcs - parquet files stored on GCS
  • bq - parquet files stored on GCS and a BQ dataset with tables linked to GCS files.
• Refactors main logic to be more DRY considering the addition of endpoint.

closes #73

[dkapitan]

Looks good to me!