Implement bulk generation module

[AusIV]

I needed to generate several million names for a sample dataset,
and looking up names out of the file every time was very time
consuming.

I left the original logic untouched, as it's the best approach for
a quick, one-off lookup. I've added a separate module, 'names.bulk'
which offers the same interface, but caches the entire file in memory
and picks names with a binary search instead of a scan.

Should resolve: #3

[coveralls]

Coverage decreased (-1.3%) to 98.75% when pulling b0eb4b4 on AusIV:master into c485a43 on treyhunner:master.

[coveralls]

Coverage remained the same at 100.0% when pulling 79934f8 on AusIV:master into c485a43 on treyhunner:master.

[coveralls]

Coverage remained the same at 100.0% when pulling 495e650 on AusIV:master into c485a43 on treyhunner:master.

[AusIV]

I'm a bit perplexed. The Travis-CI build seems to be failing on Python 3.2 on something that has nothing to do with my changes. It looks like the coverage package is failing on Python 3.2

[coveralls]

Coverage remained the same at 100.0% when pulling 808e1b0 on AusIV:master into c485a43 on treyhunner:master.

[dblackdblack]

@treyhunner this looks like a great PR

[altendky]

So I failed to bother to look at PRs before writing my own ([WIP]) in #23 which is 'similar' to this. Here at least a bisect is used but I think the random.choices() with weights specified might be better? I don't know how they actually implement it but maybe something to consider.

[alexisszabo]

@treyhunner Any chance we could approve this PR? This looks like a big performance boost on apps that call the library repeatedly. Anything I could do to help if QA needed?