Remove unsupported Lance dataset option (#1)

CI fails against pylance 8.0.0 because `lance.dataset` no longer accepts
`ignore_namespace_table_storage_options`. This removes that argument
from the TensorFlow adapter API and keeps namespace table lookup on the
supported `namespace_client` and `table_id` arguments.

Local validation:
- `uv run ruff format --check .`
- `uv run ruff check .`
- `uv run pytest -q` (TensorFlow tests skipped locally because
TensorFlow is not installed on this macOS environment)

Author: Xuanwo

src/lance_tensorflow/data.py

@@ -156,7 +156,6 @@ def from_lance(
     output_signature: Optional[Dict[str, tf.TypeSpec]] = None,
     namespace_client: Optional["LanceNamespace"] = None,
     table_id: Optional[List[str]] = None,
-    ignore_namespace_table_storage_options: bool = False,
 ) -> tf.data.Dataset:
     """Create a ``tf.data.Dataset`` from a Lance dataset.
 
@@ -184,9 +183,6 @@ def from_lance(
     table_id : Optional[List[str]], optional
         Table identifier used together with ``namespace_client`` to locate
         the table.
-    ignore_namespace_table_storage_options : bool, default False
-        When using ``namespace_client``/``table_id``, ignore storage options
-        returned by the namespace.
 
     Examples
     --------
@@ -237,7 +233,6 @@ def from_lance(
             dataset,
             namespace_client=namespace_client,
             table_id=table_id,
-            ignore_namespace_table_storage_options=ignore_namespace_table_storage_options,
         )
 
     if isinstance(fragments, tf.data.Dataset):

tests/test_data.py

@@ -119,19 +119,17 @@ def fake_dataset(uri=None, **kwargs):
         None,
         namespace_client=ns,
         table_id=["tbl"],
-        ignore_namespace_table_storage_options=True,
     )
 
     assert calls["kwargs"]["namespace_client"] is ns
     assert calls["kwargs"]["table_id"] == ["tbl"]
-    assert calls["kwargs"]["ignore_namespace_table_storage_options"] is True
 
     batches = list(ds)
     assert [b["a"].numpy().tolist() for b in batches] == [[1, 2]]
 
 
 def test_scan_use_tf_data(tf_dataset):
-    ds = tf.data.Dataset.from_lance(tf_dataset)
+    ds = tf.data.Dataset.from_lance(tf_dataset, batch_size=100)
     for idx, batch in enumerate(ds):
         assert batch["a"].numpy()[0] == idx * 100
         assert batch["s"].numpy()[0] == f"val-{idx * 100}".encode("utf-8")