Update word count to match canonical tests

[matthewmorgan]

Closes #445

The regex is not my favorite, but it gets the job done. Also, the tests now match the canonical data.

I also took the liberty of skipping all but the first test. If #518 gets merged, these tests will run in CI ;)

[behrtam]

This quote stuff is somehow overcomplicating this otherwise simple exercise. I agree that a regex is the best solution in this case ... otherwise one would have to write a lot just to handle the special cases.

[matthewmorgan]

@behrtam yes, the quote thing is thorny. Solving the exercise was trivial except for that.

[ilya-khadykin]

While implementing this (or rather improving) example solution for this exercise for bash track I simply ignored the case with ' from canonical-data.json
So, it isn't necessary to include all the tests, but it's generally a good idea

Thanks for the PR, @matthewmorgan