uwazi
uwazi copied to clipboard
New Thesauri UI
Implement the new Thesauri UI as in the design
- Can top-level thesaurus be dragged into a group? What happens to denormalization processes?
Pretty much what we've been doing in most settings pages with the new empty state for tables that should propagate to any table that could find itself empty (No [x]).
Settings / Thesauri - empty
Settings / Thesauri - list
Settings / Thesauri - selected (+action-bar)
Settings / Thesaurus - new (empty)
Settings / Thesaurus - add item
Settings / Thesaurus - list
Settings / Thesaurus - edit item
Settings / Thesaurus - add group
Settings / Thesaurus - list - add group + item
Settings / Thesaurus - list - group expand
Settings / Thesaurus - list - selected (+action-bar) item & group expand
If moving items inside and outside is possible: we could add bulk actions to group and ungroup (those buttons are not required for this iteration)
**There's a slight change that I'm proposing only for the table actions/buttons, so they become less present ("in-your-face"), but only if possible. :)
The current Thesauri UI is on charge of the Topic Classification feature!
This feature automatically suggests dictionary options for entities. The suggestions are handled using two specific windows within Uwazi's Thesauri UI. To access these windows, there are some buttons in the Thesauri UI.
Topic classification is behind a feature toggle that is activated for each object in the dictionary collection, using the following field:
enable_classification: true
For each of the dictionaries that has topic classification enabled, a button appears in the UI.
And when the new button is pressed, it redirects to "/settings/dictionaries/cockpit/{id}"
@gabriel-piles this is unrelated to what we are developing right now. This is meant for topic classification for paragraphs that are contained in a rich text field. Before supporting this we first need to have another feature that is a paragraph extractor and another feature that is ML over the rich text field contents.
Also, I would like no to pollute the thesauri UI with anything related to ML.
So let's ignore this for now.