upgrade hightable to 0.13.1

[severo]

For reference, here is the CHANGELOG for hightable 0.13.0: https://github.com/hyparam/hightable/blob/master/CHANGELOG.md#0130---2025-03-21

• adapt styles
• handle the new orderBy argument in parquet dataframe, workers and view (descending order + multiple columns)

This PR was not trivial because it required handling multi-column sorting for parquet.

I implemented it with the following details:

• I created a new function, getParquetColumn, to fetch the values of a given column without transposing to rows (using parquetRead and onChunk)
• I sorted the values of a column, stored the sorted indexes, then computed the ranks instead (so that two matching values have the same rank, which is required to implement multi-column sort), and finally cached these ranks for a given column
• I computed and cached the indexes for a given orderBy criteria by getting the ranks for each column and sorting the rows using the following column for each tie.

[severo]

Also, note that for parquet, as we're interested in getting the ranks of a sorted column, not the values, we could improve the performance by using the row group metadata (min/max). But I guess it's more something to be done in hyparquet.