explorer block list

[Don-Isdale]

In the explorer, the block is the unit of action : the user clicks on a block to add it to the display. Currently blocks are grouped into datasets, which is not the most useful organisation when selecting blocks to add to the display, because the user will be interested in selecting blocks of the same scope from multiple datasets, and would be interested in datasets organised by parent. . This task is to show an alternate view of the explorer list / tree, without the dataset level. There are 2 GUI designs discussed for presenting the block list : a flat list and a 3-level tree (parent : scope : block). For both solutions, the blocks will be grouped by their parent and scope. If the flat list is significantly easier to implement then it may be done first as an initial step.

List of blocks

Display the block list as a flat list, with the scope & dataset appended to the name (or in 2 additional columns) to make the block name unique (block name is normally the scope). The parent name could be used instead of the dataset name, but dataset name is more informative and each dataset may have 1 parent, so using dataset name here is sufficient to stand in for parent name. This could be presented as a table, with separate columns for parent, dataset and scope. This would have the advantage that we can use a table module such as HandsOnTable which has filtering / sorting by columns. This view may be useful as an alternative of the tree view.

Block tree browser

Like the current dataset explorer, but with 3 levels : parent : scope : block. If there are >1 blocks in each parent:scope, then the list of blocks can be initially collapsed into a '+' as is currently done in the dataset explorer. Similarly, we could collapse scopes into each parent, so the initial display would be just the parents.

It is suggested that the initial explorer view before the user applies filtering will be the current dataset explorer, then when filtering is applied, the view will switch to the block list or block tree browser.\ Filtering is described in #105.

'Manage Data' panel

There are some operations on datasets which are currently accessed from the dataset explorer, e.g. changing permissions, delete, and renaming. These can be surfaced in the manage data panel (currently labelled 'Mine').

---

part of #57

[josquint]

This also relates to Nested display of child datasets in Explorer #67

[kieranongh]

I've made a mockup of what I think is required

Is this what it should like? (Parent -> Scope -> Block (labeled by dataset)

Just realised I forgot to add one for the parent

Maybe with a better distinction between levels (parent, scope, block)