[kafka] Transcode mapped raw and string records to Arrow

### Purpose

Part of #4185. PR06 of the Kafka compatibility sequence.

Use the DDL mapping contract to decode raw/string keys and values, assemble physical rows with optional timestamps and ordered headers, and encode owned native Arrow log bytes.

### Acceptance criteria

- [ ] Raw bytes and strict UTF-8 strings map to the configured physical columns.
- [ ] Nullable values, empty bytes, NOT NULL violations and duplicate/nullable headers are handled correctly.
- [ ] The native batch uses the table schema ID and remains readable after Arrow resources close.
- [ ] Arrow writers and allocators close on encoding failures.

### Dependencies

PR05 Produce protocol (#4281); reuses the PR03 DDL contract (#4277 / #4275).

### Scope boundary

No JSON decoding, gateway append, protocol registration or core module changes.

### Implementation

Draft PR #4285: https://github.com/apache/fluss/pull/4285

Java 11 validation: 75 unit tests and 2 integration tests passed at `503136295d6bbaed81b7e7845dceb2d64db2d9d4`. Review and merge remain pending.

Start by reviewing draft PR #4285, the DDL mapping contract from #4277/#4275, and Produce protocol PR #4281. Check the 75 unit and 2 integration tests against the acceptance criteria, including nullable values, headers, schema IDs, resource cleanup, and readability after Arrow resources close.

Written by the indexing model from the issue text.