Support data sequence numbers in REST scan tasks

REST-planned `FileScanTask` currently parses `first-row-id` from the embedded `data-file` JSON, but the REST scan-task payload does not expose the manifest-entry data sequence number.

For v3 row lineage readers, `_last_updated_sequence_number` inheritance needs the data file data sequence number. Without it, REST-planned tasks can pass `first_row_id` but cannot inherit `_last_updated_sequence_number` for files that omit or store null row-lineage columns.

Follow-up from apache/iceberg-cpp#822. The PR adds a skipped test, `FileScanTasksFromJsonTest.RowLineageSequence`, that should be enabled once REST scan-task JSON can carry this value.

Start by tracing REST scan-task JSON parsing and how the manifest-entry data sequence number is represented. Use the skipped FileScanTasksFromJsonTest.RowLineageSequence test as the acceptance check. Done means REST-planned tasks carry the data sequence number and the test is enabled and passes for row-lineage inheritance.

Written by the indexing model from the issue text.