Improve dataset search/sorting

[daniel-wer]

Datasets often have quite similar names (e.g., ROI_2017_x, ROI_2017, ROI_2017_x_y), which makes it hard to find a specific dataset. Currently while searching the result is sorted by the default column (creation date), so when searching for the original ROI_2017 dataset (which was created first), the result is hidden somewhere in the last search results. The dataset ranking/sorting while searching should be improved by using a metric such as shortest match or levenshtein distance.