SPEC7: remove included code from page in favour of pointing to

[bsipocz]

The implementation code ideally should not be part of the spec as we should not go through endorser buy-in cycles for the code maintenance and other trivial changes.

E.g. #397 should not affect if a project is endorsing the concept of using Generators instead of rng keyword seeding of random numbers.

So my proposal is to move the code out to a specs tools like we did for a few other SPECs.

[jorenham]

The code in question is only used internally by SciPy, and not at all part of the public API. So I can see how it could be confusing that it's presented in the spec this way.

[mdhaber]

It's OK for it to be moved somewhere. It was good for us to write the code while writing the SPEC to work out the kinks and make it easier for projects to transition, though.