Add capability to sample source datasets by percent or rowcount to increase speed of testing

When making adjustments to CytoTable configuration alongside a source dataset it can take a large amount of time to process the data only to find the configuration fails for a reason that might be quickly detected using a smaller portion of the source dataset (for instance, using 1% of the data from the source). This issue describes a feature where one would be able to limit the data used by CytoTable using, for example the [`SAMPLE`](https://duckdb.org/docs/sql/query_syntax/sample) feature from DuckDB. This could be used to limit the data extracted to a smaller amount.