r.proj: Create unit tests

[cyliang368]

This PR creates unit tests for the r.proj. The results are checked by the extents, nrows, ncols, and statistics of the reprojected raster.

Closes #1663

[petrasovaa]

Should we close #1663?

[echoix]

Should we close #1663?

Does it contain some useful parts that could be taken here?

[echoix]

Should we close #1663?

Does it contain some useful parts that could be taken here?

At first glance, I see that #1663 has more fields of the statistics in the expected results. Like sum, mean_of_abs, coeff_var, range. I don't see them mentioned in the current docs of r.proj, was it included at some point?

[cyliang368]

Should we close #1663?

Does it contain some useful parts that could be taken here?

At first glance, I see that #1663 has more fields of the statistics in the expected results. Like sum, mean_of_abs, coeff_var, range. I don't see them mentioned in the current docs of r.proj, was it included at some point?

#1663 has everything from r.univar, but this is not necessary. For example, if max and min are determined, range will also be determined. mean can cover information for sum and mean_of_abs.

[echoix]

Ok, I see now how it's univar. In that case yes, I agree that the rest was kinda redundant. Like variance and stddev. So we can close the older PR then

All changes have been addressed

[echoix]

I misspressed on update instead of merge (very slow network), so we'll need to wait..