feat: Collection View Improvements

[chriswhong]

Status Quo

Collection view is currently an alphabetical list of dataset references in the user's local repo.

• User can filter the list using local string matching
• User can click a dataset to load it into workbench view
• User can create a new dataset using button (opens modal) or drag and drop file
• User can clone a dataset using button (opens modal)
• The list lives in the left pane. The right pane is unused except as a drop zone.

Improvements (Needs Mockups)

Multi-select list view(s)

Collection should be a table view, more akin to a folder listing that shows name and descriptive info.

• User can sort the list by any of the headers (username, dataset name, last commit timestamp, size, rows, etc.
  • dataset username
  • dataset name
  • latest commit
  • size
  • entries
  • published/unpublished
  • up to date/needs sync
• User can select one or many datasets and perform actions on those datasets
  • Remove
  • Duplicate
  • Publish/Unpublish
  • Sync

Group by username vs Consolidated List

• User can toggle between a consolidated list of all datasets in their repo or a set of lists for each username. The local user's list will be shown first, other users' datasets will follow.

• Or, should there be a notion of "my datasets" vs "other users' datasets" and we show two lists instead of one for each user?

Recently Used Datasets

• Add a pane of 3 most recently used datasets. When a user returns to collection view, they can quickly re-open datasets they recently had open.

Cloning and "new" indicator

Cloning currently opens the newly-cloned dataset in the workbench. Cloning from collection view should show a progress bar and result in a new item added to the list. Use a colored dot or glyph to show "new" datasets that have not been viewed (or maybe leave the new indicator in place for a set time, perhaps 24 hours)

Importing

Importing a CSV should behave in a similar way to cloning, show a progress bar, add the item to the list with a new indicator.

Filtering

Sorting may be enough to quickly find datasets in the list, but filtering or some kind of search may be needed when collections get larger. Th