Feature: Maintain link between kedro data catalog and .dvc files

[shaunc]

In order to track data used in kedro pipelines, we need to maintain a link between the kedro data catalog and data tracked by dvc.

See discussion.

For each supported entry in catalog, we create .dvc as described in discussion. We

Cases:

• where is configuration: default or configured in different location?
• kedro environment present in
  • base
  • local
  • test
• dvc status
  • no .dvc
  • .dvc already exists
• whether we run via command line or in hook
  • command line
  • on hook after_catalog_created:
    • same as kedro dvc update without --force

Scenarios:

• kedro dvc update
  • if not exists, create
  • if exists, and corresponds, do nothing
  • if exists and different, error
• kedro dvc update --force
  • same as above, except if exists and different, overwrite
• environment besides base.

Tasks:

• data catalog <-> .dvc files: setup command: kedro dvc install/update #33
• data catalog <-> .dvc files: one file in catalog.yml - no existing .dvc #34
• data catalog <-> .dvc files: one file - existing dvc #35
• data catalog <-> .dvc files: one file - existing dvc, catalog target changes -- fail w/o --force #37
• data catalog <-> .dvc files: one file - existing dvc, catalog target changes #36
• data catalog <-> .dvc files: case different conf location #38
• data catalog <-> .dvc files: add cases: different kedro environments #39
• data catalog <->.dvc files: two entries in catalog #40
• data catalog <-> .dvc files: add case: run in API vs not #41