Conversion methods: to* versus to_*?

[dmdunla]

(I was going to make a Shakespeare pun, but I resisted...)

What should the convention be for the conversion methods?

Here are a couple in place already:

• sptensor.spmatrix (aligns with MATLAB TTB)
• tensor.to_sptensor
• ktensor.tolist
• ktensor.tovec
• ktensor.to_tensor

Note: numpy and scipy support to* methods. Examples (using tab completion to illustrate):

>>> np.ndarray.to
np.ndarray.tobytes(   np.ndarray.tofile(    np.ndarray.tolist(    np.ndarray.tostring(  
>>> sp.coo_matrix.to
sp.coo_matrix.toarray(  sp.coo_matrix.tobsr(    sp.coo_matrix.tocoo(    sp.coo_matrix.tocsc(    sp.coo_matrix.tocsr(    sp.coo_matrix.todense(  sp.coo_matrix.todia(    sp.coo_matrix.todok(    sp.coo_matrix.tolil(

Suggestion: In pyttb, we should align with numpy and scipy conventions.

@jeremy-myers @ntjohnson1 Please weigh in.

[ntjohnson1]

I find to_* more readable and I think it follows the snake case convention. However I think aligning with numpy is probably the better option at this point.

[jeremy-myers]

I agree with @ntjohnson1 that to_* is more readable and follows snake case convention. I am less experienced and more prone to statements like "numpy and scipy got to* wrong".