Add UnavailableDataset as a default dataset for lite mode

[ravi-kumar-pilla]

Description

Related to #2058

Development notes

• Created a custom dataset UnavailableDataset
• Updated the default dataset from MemoryDataset to UnavailableDataset in --lite mode.
• Updated filepath to have an empty string as N/A instead of - as suggested by @stephkaiser
• Updated tests

QA notes

• When you run kedro viz --lite with kedro-datasets missing in the conda env, the metadata panel for the dataset should show the dataset type as UnavailableDataset if the dataset is mentioned in catalog.py (i.e., not a MemoryDataset)

• All tests should pass

Checklist

• Read the contributing guidelines
• Opened this PR as a 'Draft Pull Request' if it is work-in-progress
• Updated the documentation to reflect the code changes
• Added new entries to the RELEASE.md file
• Added tests to cover my changes

[astrojuanlu]

Thanks @ravi-kumar-pilla - could you clarify what was the situation before this PR?

[ravi-kumar-pilla]

Thanks @ravi-kumar-pilla - could you clarify what was the situation before this PR?

Hi @astrojuanlu , Thanks for reviewing. Before this change, the default was MemoryDataset . This PR is part of making users aware of the missing dependency. MemoryDataset will be a bit misleading as the dataset is registered in the catalog.

[stephkaiser]

thanks Ravi! using Gitpod, filepath is showing as N/A for me 👍 but still shows MemoryDataset, maybe i did something wrong?

[ravi-kumar-pilla]

thanks Ravi! using Gitpod, filepath is showing as N/A for me 👍 but still shows MemoryDataset, maybe i did something wrong?

Thanks for reviewing. Yes these are actually MemoryDatasets (as they are not registered in the DataCatalog). We will show UnavailableDataset to the datasets which are registered in the DataCatalog but the dependency is not installed.

For example, companies is registered in DataCatalog as below but has a missing dependency of pandas.CSVDataset, so companies metadata will have dataset type as UnavailableDataset. I hope this is clear. Thank you

companies:
  type: pandas.CSVDataset
  filepath: ${_base_location}/01_raw/companies.csv
  metadata:
    kedro-viz:
      layer: raw
      preview_args:
        nrows: 5

[stephkaiser]

@ravi-kumar-pilla understood, thank you!

[rashidakanchwala]

LGTM. thanks! <3