Handle zero-sized ndarray more gracefully

[huxuan]

Environment

• OS: [Ubuntu 22.04]
• Hardware (GPU, or instance type): [H800]

To reproduce

Steps to reproduce the behavior:

1. Create a python script (e.g. main.py) with the following snippets:

   from streaming import MDSWriter
   import numpy as np
   
   with MDSWriter(out="data/", columns={"data": "ndarray"}) as writer:
      writer.write({"data": np.empty(())})

2. run with python main.py and it will complain about

   Traceback (most recent call last):
    File "/Users/huxuan/Code/test/1.py", line 5, in <module>
      writer.write({"data": np.empty(())})
    File "/Users/huxuan/Code/test/.venv/lib/python3.12/site-packages/streaming/base/format/base/writer.py", line 259, in write
      new_sample = self.encode_sample(sample)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/Users/huxuan/Code/test/.venv/lib/python3.12/site-packages/streaming/base/format/mds/writer.py", line 106, in encode_sample
      datum = mds_encode(encoding, value)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/Users/huxuan/Code/test/.venv/lib/python3.12/site-packages/streaming/base/format/mds/encodings.py", line 619, in mds_encode
      return coder.encode(obj)
             ^^^^^^^^^^^^^^^^^
    File "/Users/huxuan/Code/test/.venv/lib/python3.12/site-packages/streaming/base/format/mds/encodings.py", line 244, in encode
      shape_dtype = self._rightsize_shape_dtype(shape_arr)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/Users/huxuan/Code/test/.venv/lib/python3.12/site-packages/streaming/base/format/mds/encodings.py", line 204, in _rightsize_shape_dtype
      if shape.min() <= 0:
         ^^^^^^^^^^^
    File "/Users/huxuan/Code/test/.venv/lib/python3.12/site-packages/numpy/core/_methods.py", line 45, in _amin
      return umr_minimum(a, axis, None, out, keepdims, initial, where)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
   ValueError: zero-size array to reduction operation minimum which has no identity

Expected behavior

I think we should either handle the case silently or give more friendly error message.

Additional context

[snarayan21]

Hey @huxuan, thanks for raising this issue! If you wouldn't mind, could you open up a PR that adds a better error message (ValueError) in this ndarray encoding function for when the input np array has no elements? We always welcome open source contributions!