WKB import rejects LargeBinary storage

## Summary

Importing WKB whose Arrow storage is `LargeBinary` fails:

```
SedonaDB internal error: Expected Wkb type with Binary storage but got LargeBinary.
```

Reproducer through a literal (the same applies to any LargeBinary-backed `geoarrow.wkb` input):

```python
import geoarrow.pyarrow as ga, pyarrow as pa
from shapely.geometry import Point
from sedonadb.expr import lit

df.select(lit(pa.array([Point(1, 1).wkb], ga.large_wkb())))
```

`ga.large_wkb()` is a public GeoArrow type, and sources such as Iceberg return BLOB columns as LargeBinary, so this shows up with real data, not only constructed literals.

## Expected

WKB with `LargeBinary` storage is accepted wherever `Binary`-backed WKB is, keeping the CRS and edge-type metadata of the extension type.

## Related

- #1195 works around this on the Python side by rebuilding the type on `Binary` storage (with the same CRS and edge type) before resolving the literal; that can go once the importer accepts LargeBinary.

Start by running the Python reproducer through the lit entry point and tracing the WKB importer error for LargeBinary storage. Confirm that LargeBinary-backed WKB is accepted while preserving CRS and edge-type metadata, then verify that the workaround described in #1195 is no longer needed.

Written by the indexing model from the issue text.