factor out license patterns

[maxheld83]

part of path was missing & licence_patterns is needed as a vector

[maxheld83]

csv file from https://github.com/subugoe/hoad/commit/3d8b8f4a69b022812ecbe59bad66df445e105c4d is still missing for this PR to work.

(Apologies @jhoeffler, this was my mistake, I think my reverts on #172 killed the csv, but I'll carry it over in the merge).

[codecov-commenter]

Codecov Report

Merging #213 into master will not change coverage. The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #213   +/-   ##
=======================================
  Coverage   62.22%   62.22%           
=======================================
  Files           4        4           
  Lines          45       45           
=======================================
  Hits           28       28           
  Misses         17       17           

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 52b3ffe...8e3cc1e. Read the comment docs.

[maxheld83]

Ok, I have now refactored @jhoeffler refactor 🙃 as a properly documented and tested R object inside the package. This, as I have been suggesting for some time, should, as far as possible, be the modus operandi going forward for all new additions/refactors/etc. Slowly but steadily, we should be factoring out ever more stuff from monolithic functions and scripts into such small, tested and documented building blocks.

This seems a bit weird in this case, because the object is just a tribble, and doesn't really do much. Though considering how much debate there was around this and how many comments were in the script including sources, etc, I think this is how we should be moving forward even for relatively minor components.

In this case, I:

• created licenses.R which just creates the object
• I've included the source in the object, and details on the source in the objects documentation.
• the object is documented and exported
• the object is also tested, currently quite perfunctorily, only whether the URLs are valid, but more to come perhaps in #215.

This is just a dress rehearsal for the real deal in #47 and #81.

Unfortunately, the above diff is a bit messed up, because there were:

• auto-styling changes (RStudio, anyone?)
• whitespace changes (it's easy to disable them in your local git config, but you can also just toggle them off in any diff views).

A short note on where to store such things: inst/extadata/*.csv have some disadvantages:

• the path changes at run time, so some care must be taken with using it, and there are some gotchas during development (i.e. devtools has to shim the path for devtools::load_all() to work.).
• *.csvs are inherently undertermined compared to R objects: readr::read_csv() makes a bunch of assumptions (such as on whether something is a factor or a character vector), and these might be wrong or change over time. So to be safe, one would have to explicate all of these in the readr::read_csv() call, which ends up being a lot of work. An alternative is to just store these objects in a tribble, which is a tibble conveniently written row-wise (and therefore, easier to read). There's no extra care needed here, because it's just a proper R object.
• In this case, I think we really also wanted comments inside the data objects, and CSVs don't do that.

I'm going to leave this open, to wait for @jhoeffler's review, and will then merge it if it's allright.