Support ASGI async export streams, refactor XLSX payload, add importer bulk-update and tests

[TimurKady]

Motivation

• Enable non-blocking exports under ASGI by providing asynchronous stream generators to avoid Django warnings about synchronous iterators.
• Provide a thread-safe way to build XLSX payloads so large-file exports can be produced for both sync and async handlers.
• Allow the importer to update existing rows (by primary key) during import instead of only creating new records.
• Add regression tests to validate async export streaming and importer update behavior.

Description

• Added asgiref.sync.sync_to_async usage and helper TreeNodeExporter._get_sync_iterator_item and async stream variants async_csv_stream_data, async_json_stream_data, async_yaml_stream_data, and async_xlsx_stream_data to treenode/admin/exporter.py.
• Refactored XLSX generation into _build_xlsx_payload and adjusted xlsx_stream_data to yield the built payload, and added _resolve_stream_data to select async or sync stream plus process_record(request=None) to detect ASGI requests.
• Updated treenode/admin/mixin.py to pass the request into process_record so the exporter can choose async streams for ASGI.
• Enhanced treenode/admin/importer.py to detect existing objects per level via _get_existing_objects_map, normalize parent references with _normalize_parent_reference, collect to_update items, and perform bulk_update using _get_update_fields, while tracking result['updated'].
• Extended treenode/tests.py with AdminExportAsyncTests (async export streaming) and TreeNodeImporterTests (import updates existing row), and added necessary imports and fixtures.

Testing

• Added unit tests in treenode/tests.py including test_export_endpoint_uses_async_stream_without_warning and test_import_tree_updates_existing_row_by_pk.
• Ran the Django test suite (test cases under treenode.tests) exercising async export and importer update paths.
• All automated tests completed successfully.

---

Codex Task