Support data files compaction

Introduce an API to compact data files. The first version of the API will do the following:
- take a predicate expression as input parameter to find data files matching the filter that will be re-written
- group data files by partitions and rewrite them using the same bin-packing constraints of the writer

Start by reading PyIceberg's existing data-file and writer APIs to understand how predicate filtering, partition grouping, and bin-packing constraints are represented. Trace the writer path and relevant tests before deciding where the compaction API belongs. Done means callers can provide a predicate, matching files are grouped by partition, and rewritten files use the writer's existing bin-packing constraints.

Written by the indexing model from the issue text.