Adjust minimal presentation output to ensure no extra x-grades and y-grades are included

[mlesnick]

The output for the minimal presentation computation is handled as follows:

1. Early in the computational pipeline, the unique x-grades and y-grades of the FI-Rep are printed as lists X and Y.

2. After the minimal presentation is computed, the bigrades for each row and column are printed using an integer representation, where (X(i),Y(j)) is represented as (i,j).

3. The matrix is printed in a column-sparse format.

Usually, many of the elements of X and Y will not be grades of generators or relations for the minimal presentation. So what we are outputting is usually not truly minimal. It is a rather superficial issue, but should be addressed.

This is related to issue #99, and perhaps both should be addressed at the same time.