build(datasets): Make sure experimental datasets are packaged as part of `kedro-datasets`

[merelcht]

Description

Closes #653

Development notes

Possible solutions:

Solution 1

• https://github.com/kedro-org/kedro-plugins/pull/654/commits/a17cb8a64412506745f0c39043e9be010b25c743
• To use an experimental dataset users can just do from experimental.<...> import <...>

✅ Easy to implement packaging ✅ Position of the experimental directory makes it very clear that the datasets inside it aren't part of the core datasets ❌ experimental is a very generic name and it wouldn't be clear from the import that this is Kedro related

Solution 2

• https://github.com/kedro-org/kedro-plugins/pull/654/commits/aab299b14b8e440fe97f153944ada2925a478304
• Rename the experimental directory to kedro_datasets_experimental
• The import will then become from kedro_datasets_experimental.<...> import <...>

✅ Easy to implement packaging ✅ Position of the kedro_datasets_experimental directory makes it very clear that the datasets inside it aren't part of the core datasets ✅ Kedro is in the name of the import

Solution 3

• Move the experimental directory inside the kedro_datasets directory
• The import will then become from kedro_datasets.experimental.<...> import <...>

✅ Easy to implement ✅ Kedro is in the name of the import ❌ Position of the experimental directory makes it less clear that this is not just a core dataset, but the directory that contains the experimental datasets

Solution 4

• Turn kedro_datasets and experimental into namespaced packages: https://packaging.python.org/en/latest/guides/packaging-namespace-packages/

✅ Very clear separation ❌ Harder to implement and also requires more setup for releasing both packages.

Checklist

• [ ] Opened this PR as a 'Draft Pull Request' if it is work-in-progress
• [ ] Updated the documentation to reflect the code changes
• [ ] Added a description of this change in the relevant RELEASE.md file
• [ ] Added tests to cover my changes