Lists

Most of your time on Picsellia starts from a list: the Dataset of the Organization, the versions of a Dataset, the Project, the Model, the Deployment, the Processing. Every list of the platform follows the same layout and offers the same tools, so once you know how to find, filter, and act on the objects of one list, you know how to do it everywhere.

1. The layout of a list

From top to bottom, the header of a list always holds the same elements:

• the display mode toggle, Grid or Table, followed by the search bar, which filters the list by name;
• the filters: the Order by field with its ascending / descending toggle, the Select Tags filter, and, on some lists, filters of their own, such as the Version Type of the versions of a Dataset;
• the actions on the selection, displayed once you tick one or more objects.

The button to create a new object, such as + New Dataset, sits at the top right of the page.

2. Choose how a list is displayed

The Grid mode displays each object as a card with its preview, which suits the lists where you recognize objects by their images, such as Dataset. The Table mode displays one row per object with more columns, such as the number of versions, the creation date, and the members.

Your choice is remembered for each list separately: if you browse the Dataset as a grid and the Model as a table, each list opens in its own mode the next time.

Lists load their objects as you scroll instead of page by page, so even a list of several hundred thousand items stays fast to browse.

3. Act on a selection

Tick the checkbox of one or more objects: the number of selected objects is displayed above the list, next to the actions available on them. On the lists of platform objects, the Tag action attaches or detaches tags, as detailed in Tags.

The lists of Data, Asset, and PredictedAsset offer more actions on the selection, detailed in Datalake overview, Assets overview, and Predictions overview.

4. Write descriptions in Markdown

The description of an object, for instance in the Settings of a Dataset, accepts Markdown. It is the place to record what a list cannot tell at a glance: where the images come from, which guidelines they were annotated with, or what a version should not be used for.

The description field has two tabs: Write, to edit the text, and Preview, to check how it renders. The toolbar of the Write tab inserts a heading, bold, italic, a quote, code, a link, a bulleted list, a numbered list, and a task list, and each of them has a keyboard shortcut:

Once saved, the description is rendered wherever it is displayed.