`len` of points (used in `sdata.__repr__`) is broken after dask unpin

[LucaMarconato]

Instead of a nice number for the length (or a placeholder for when we don't know the length), the repr of sdata shows this for points:

SpatialData object, with associated Zarr store: /Users/macbook/embl/projects/basel/3d-spatial-workshop-2025/data/merfish_mouse_ileum.sdata.zarr
├── Images
│     └── 'stains': DataTree[czyx] (2, 9, 9392, 5721), (2, 4, 4696, 2860), (2, 2, 2348, 1430)
├── Labels
│     ├── 'dapi_labels': DataTree[zyx] (9, 9392, 5721), (4, 4696, 2860), (2, 2348, 1430)
│     └── 'membrane_labels': DataTree[zyx] (9, 9392, 5721), (4, 4696, 2860), (2, 2348, 1430)
├── Points
│     ├── 'cells_centroids_cellpose': DataFrame with shape: (<dask_expr.expr.Scalar: expr=df.size() // 3, dtype=int64>, 3) (3D points)
│     ├── 'molecule_baysor': DataFrame with shape: (<dask_expr.expr.Scalar: expr=(AssignAlign(frame=AssignAlign(frame=Assign(frame=(AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=df, column='gene', value=df), column='nuclei_probs', value=df), column='is_noise', value=df), column='area', value=df), column='assignment_confidence', value=df), column='total_magnitude', value=df), column='ncv_color', value=df), column='mol_id', value=df), column='x_raw', value=df), column='z_raw', value=df), column='brightness', value=df), column='y_raw', value=df), column='molecule_id', value=df), column='cell', value=df), column='compartment', value=df), column='qc_score', value=df), column='confidence', value=df))[['mol_id', 'x_raw', 'y_raw', 'z_raw', 'gene', 'area', 'brightness', 'total_magnitude', 'qc_score', 'x', 'y', 'z', 'molecule_id', 'confidence', 'compartment', 'nuclei_probs', 'cell', 'assignment_confidence', 'is_noise', 'ncv_color']]), column='layer', value=df), column='layer', value=df)).size() // 21, dtype=int64>, 21) (3D points)
│     └── 'molecules': DataFrame with shape: (<dask_expr.expr.Scalar: expr=(AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=AssignAlign(frame=df, column='gene', value=df), column='area', value=df), column='total_magnitude', value=df), column='z_um', value=df), column='brightness', value=df), column='molecule_id', value=df), column='y_um', value=df), column='x_um', value=df), column='qc_score', value=df)).size() // 12, dtype=int64>, 12) (3D points)