RevKit interface

[msoeken]

I propose to merge the RevKit feature into develop. The tests are running with 100% coverage, the code is documented, and there are two examples. I am happy to add more functionality in the future. I also thought of having some simple Python-based fallback algorithms for synthesis that can be called if RevKit is not installed but do not guarantee the same quality of results.

[thomashaener]

Awesome, thanks! It looks great except for a few pep8 issues (which I could fix if you like) and I was wondering if there is a chance we could change the writing to a file and then exec_from_file to directly running the ProjectQ code in revkit. E.g., have a function revkit.run_projectq(qs) or some other way to directly get the gates to apply without exec_from_file. If this requires too much work, we can also do this later.

[msoeken]

Thanks for the feedback. And both good ideas. I can fix the pep8 issues and will also modify the interface to avoid writing to files. I'll push the changes within a few days.

[damiansteiger]

What does this test?

[damiansteiger]

Would be great to incorporate this PR before the next release.

I fixed the pep8 issues.

A few things:

• Is there a way to exclude revkit from coveralls for Python3.5 and Python3.6?
• Unrelated to this code: Travis lately fails often with

The command "sudo -E apt-get -yq --no-install-suggests --no-install-recommends --force-yes install gcc-4.9 g++-4.9" failed and exited with 100 during .

Rerunning was so far successful in fixing this. But maybe we should check if we find a more stable solution ;)

• What else is missing besides a final code review? (Avoid writing to files as suggested above)

[msoeken]

The reimplementation somehow diverged the branches too much such that it was cleaner to open a new PR #241 and close this one.