It seems the pruning map goes the opposite way, check:

i14 : * "minimalPresentation(Module)" -- minimal presentation of a module

o14 = minimalPresentation(Module) -- minimal presentation of a module
      ***************************************************************

      Synopsis
      ========

        * Usage: N = minimalPresentation M
        * Function: "minimalPresentation"
        * Inputs:
            * M, a module
        * Consequences:
            * minimalPresentation stores the isomorphism from M to N as g = N.cache.pruningMap unless M.cache.pruningMap already exists,
              in which case N is the same as M and the inverse isomorphism is obtained by g^-1
