Adds stringsimmatrix function

[JBGruber]

Hi @markvanderloo and the other contributors,

Thanks for this great package!

I use it in one of my packages where it really solved an issue for me. However I want to re-write the function that is built on top of stringdist at some point in the future and thought about what I need for that.

A stringsimmatrix function was what came to mind and rather than putting it into my package I looked into your code and saw that it is rather easy to accomplish that.

The new function is basically all your code from stringsim and stringdistmatrix rearranged.

[coveralls]

Coverage decreased (-0.01%) to 89.056% when pulling 6874d59 on JBGruber:master into dc4157a on markvanderloo:master.

[markvanderloo]

Hi Johannes,

this looks like a great contribution, thanks! Before I merge, could you add just a few tests to inst/tinytest/test_stringsim.R? Maybe just check dimensions and compare a simple matrix with output computed 'manually' with stringsim.

I know the code is quite short, but adding tests will give some protection if we somehow need to update one of the functions that is used by stringsim.

Best,
Mark

[JBGruber]

Hi Mark,

Sure! I wasn't familiar with your tinytest package and wanted to wait if there is a general interest in the PR before looking into how to write some tests. But it's actually very straightforward.

I used a character vector from a built-in dataset (islands) to calculate some real similarities. I think that makes sense in this case but I can also copy your approach from testing stringdistmatrix (I only had the idea after already writing the other tests). Let me know what you prefer.

I don't know why coveralls stills shows that coverage decreases. As far as I can see all my code is covered.

Best
Johannes

[markvanderloo]

Thanks! accepted.