Flink: Lookup join support

### Feature Request / Improvement

I propose to resume work on the Lookup Join functionality for flink engine (https://github.com/apache/iceberg/pull/15183).
Ready to contribute based on closed PR and according the [last comment](https://github.com/apache/iceberg/pull/15183#pullrequestreview-3727232366)

### Motivation

Enable direct, efficient joins between high-throughput streams and slowly-changing dimension data stored in Iceberg tables.

### Implementation proposal

The Lookup Join implementation will be through the IcebergTableSourceFactory factory and the IcebergTableSource class, which create the IcebergLookupFunction search function. The main architecture includes IcebergLookupFunction for performing point queries on data and an optional IcebergLookupCache component for optimizing input-output.

### Query engine

Flink

### Willingness to contribute

- [x] I can contribute this improvement/feature independently
- [ ] I would be willing to contribute this improvement/feature with guidance from the Iceberg community
- [ ] I cannot contribute this improvement/feature at this time

Start by reviewing closed PR #15183 and its last review comment, then inspect IcebergTableSourceFactory, IcebergTableSource, and the proposed IcebergLookupFunction architecture. Determine how the optional IcebergLookupCache fits the design. Done means Flink lookup joins can perform efficient point queries against Iceberg tables.

Written by the indexing model from the issue text.