Cleanup and other various changes

[sjp]

Hi Blake -- another package another pull request.

Aside from adding in the usual Travis CI support, I've ensured that the fetchR now passes the package checker. In the process I've dropped the dependency on rgeos, as it didn't seem to be used at all.

I've also had a go had a code review. Mostly this involved the following:

• Extracting a variable to avoid repeated calculation (multiple min() and max() calls)
• Using dev.hold() and dev.flush() to speed up plot drawing -- it's probably fast enough anyway.
• Dropping isTRUE(). I've added an NA check prior to this so we can safely use if (plot) for example.
• Using anyNA(), should be faster, but mostly it's just easier to read.
• Minor formatting changes.

These changes have not been tested by myself so I can't guarantee whether they work, but they are at least syntactically correct.

A few nit-picky thoughts as an aside:

• How about something other than PNG plotting? I'm thinking PDF in particular. Might be worth separating the plotting out into a separate function (maybe using S3?) to do this for you.
• I don't know how clean your data is, but it might be worth guarding against NA or zero-length vectors. There were some places in your code where that code happen -- but it seems unlikely because you would already have run into it if that were the case.
• Exposing your data via data(). I am guessing you want to keep the data internal to the package (reasonable).

Congrats on yet another package! Keep up the good work!

[blasee]

Thanks for the tips and reviews Simon!

fetchR (unlike Rome) was built in a day so there are a few concepts I am not too happy with either like having the plotting within the same calculation function - as you point out. I'll take these considerations on board and once again thanks heaps for your contributions!