Use SQLite database in single-tenant Kamu-CLI case too

[zaychenko-sergei]

Currently we enable SQlite database in a multi-tenant workspace, when no custom database configuration is specified. The original motivation was to properly store ReBAC properties on datasets per each user to support Private Datasets epic.

However, it makes sense to enable SQlite database in a single-tenant workspace as well, as there is other data, which would greatly fit there and improve CLI operations performance:

• dataset entries
• dataset dependencies
• HEAD reference
• key metadata blocks (skip list)
• other fields that are currently stored in DatasetSummary (dataset kind, data/checkpoints size, last pull time, total number of records)