Add ability to use FieldPerp in Options

[ZedThree]

I've started work on trying to completely replace use of DataFile with OptionsNetCDF. In order to prevent that PR from being way too big, I'm trying to split it into smaller chunks. This is one necessary part that is useful even if we decide not to use OptionsNetCDF and could probably be backported?

This PR actually has a couple of other features itself:

• Add FieldPerp::getGlobalIndex, FieldPerp::setIndexFromGlobal -- moved code from DataFormat methods to FieldPerp methods
  • Names maybe not the best?
• Add Options::hasAttribute to check existence of attribute
  • This leads to slightly more readable, but less efficient, code

const auto attribute = options.attributes.find("key");
if (attribute != options.attributes.end()) {
  use(attribute->second);
} else {
  use(something_else);
}

vs

if (options.hasAttribute("key")) {
  use(options.attributes.at("key"));
} else {
  use(something_else);
}

[johnomotani]

Looks good to me, although I'm not at all familiar with OptionsNetCDF.

[ZedThree]

@bendudson Would you mind have a quick glance over the OptionsNetCDF changes?

[bendudson]

Thanks @ZedThree ! Looks good to me.

[ZedThree]

We've moved from travis-ci.org to .com, so the failed PR test was restarted, but on .com which doesn't clear the .org version... Basically all tests have passed, we've just got an old notification sticking around. This is likely to the be the case with a few open PRs