Compatibility Issue with `sparse` when building a COO matrix

[fspinna]

The new sparse==0.16.0 release raises the following error when calling to_dense:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<timed exec> in <module>

[/usr/local/lib/python3.11/dist-packages/pyrregular/accessor.py](https://localhost:8080/#) in to_dense(self, reset_time_index, ts_level, index_scale, absolute_time, concatenate_time, normalize_time)
     61     ):
     62         if reset_time_index:
---> 63             X, T = self.reset_time_index(
     64                 ts_level=ts_level,
     65                 index_scale=index_scale,

2 frames
[/usr/local/lib/python3.11/dist-packages/sparse/numba_backend/_coo/core.py](https://localhost:8080/#) in __init__(self, coords, data, shape, has_duplicates, sorted, prune, cache, fill_value, idx_dtype)
    243 
    244         if shape is None:
--> 245             raise ValueError("`shape` was not provided.")
    246 
    247         if not isinstance(shape, Iterable):

ValueError: `shape` was not provided.

This should be addressed in future updates for improved compatibility.