When adding a tag from a dataset to the query, the dataset should preserve its relative position

[arteymix]

This happens when adding a term to a category that already has a selection. As it expand the results, there could be more datasets appearing before it and the dataset will be pushed on a subsequent page.

If a term instead narrows down the result, a matching dataset in the first page will remain somewhere there, likely higher.

I'm not sure there's a simple way of avoiding this.

[arteymix]

We could keep a list of manually picked datasets and pass it to the endpoint for sorting them first, before any other datasets.

Or jump to the page that contains the dataset, but that will require being capable of telling where a dataset is.

We currently reset the page back to the first one when a filter changes. Instead of doing that, we could jump to the page containing the dataset.