MAINT: better error messages for initialization

[stsievert]

During initialization, I kept on running into an assert. This PR resolves that and provides a better error message.

Why is coords in the shape it is? I would expect coords.T to be passed to __init__. That is, instead of coords.shape == (ndim, num_points), I would expect coords.shape == (num_points, ndim).

[codecov]

Codecov Report

Merging #171 into master will increase coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #171      +/-   ##
==========================================
+ Coverage   97.12%   97.14%   +0.01%     
==========================================
  Files          11       11              
  Lines        1253     1259       +6     
==========================================
+ Hits         1217     1223       +6     
  Misses         36       36

Impacted Files	Coverage Δ
sparse/coo/core.py	96.69% <100%> (+0.06%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2dc13b5...90aa267. Read the comment docs.

[hameerabbasi]

Thanks, @stsievert! This looks great. Yeah, coords will probably become 1-D and hold just the "unwrapped" or "linearized 1-D index" soon (#164), so your question may not matter much. As to why, I'd guess it's because that's how it is in scipy.sparse.coo_matrix. It has row and col or something similar IIRC, often passed in as (row, col), and they're 1-D arrays, so here we also have the "last dimension" representing the actual coordinate and the first representing the dimension.