Partial read: Improvement to error messages

[jm9176]

Here, I am using the zarr array stored here: https://github.com/mathworks/MATLAB-support-for-Zarr-files/tree/main/test

1. Start parameter is out of bounds

>> out = zarrread('grp_v2/arr_v2',Start=[30 40])
Error using ZarrPy>readZarr (line 107)
Python Error: IndexError: Computing interval slice for dimension 0: (29, 19) do not specify a valid closed index
interval [source
locations='tensorstore/index_interval.cc:358\ntensorstore/index_interval.cc:401\ntensorstore/index_space/internal/numpy_indexing_spec.cc:499\ntensorstore/index_space/internal/numpy_indexing_spec.cc:499\ntensorstore/index_space.h:438']

Error in zarrread (line 36)
data = zarrObj.read(options.Start, options.Count, options.Stride);
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

2. Count parameter is out of bounds

>> out = zarrread('grp_v2/arr_v2',Start=[],Count=[30 40])
Error using ZarrPy>readZarr (line 107)
Python Error: ValueError: OUT_OF_RANGE: Propagated bounds [0, 20), with size=20, for dimension 0 are
incompatible with existing bounds [0, 30), with size=30. [transform='Rank 2 -> 2 index space transform:   Input
domain:     0: [0, 30)     1: [0, 40)   Output index maps:     out[0] = 0 + 1 * in[0]     out[1] = 0 + 1 * in[1]
'] [domain='{origin={0, 0}, shape={20, 25}}'] [source
locations='tensorstore/index_space/internal/propagate_bounds.cc:285\ntensorstore/index_space/index_transform.h:991']

Error in zarrread (line 36)
data = zarrObj.read(options.Start, options.Count, options.Stride);
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

3. @krisfed For the following case, we have been following to read the whole HDF5 dataset if the input is defined as inf, so would you prefer a similar behavior in case of zarr partial read as well?

If any element of count is Inf, then h5read reads until the end of the corresponding dimension.(ref)

>> zarrread('prt_grp/arr_1',Count=[inf inf])
Error using zarrread (line 31)
 zarrread('prt_grp/arr_1',Count=[inf inf])
                                ^^^^^^^^^
Invalid value for 'Count' argument. Value must be integer.

[krisfed]

Checked with Shubh: 1 and 2 are about error messages being confusing / not clear to read (especially hard to determine which parameter might be at fault). We can improve it by doing our own validation and errors for out-of-bounds conditions instead of relying on tensorstore.

3 - that was a conscious decision on my part to not introduce this additional functionality/convenience. If there is need and requests for it, we can add it later. As of now, users can skip specifying Count to read data till the end of each dimension in the Zarr array.