r4ds-style refactor

[maxheld83]

Before we continue with actual features, we need a bit of refactoring.
This refactoring has a couple of goals; they are so tightly connected that they have to be one, somewhat monolithic refactor unfortunately:

1. decoupling (this used to be refactor ELT + analysis codebase #52)
   • the ELT + analysis codebase should be organised in terms of the r4ds book
   • presentation should be separated entirely from analysis.
     The rmarkdown template (for the mail) strictly should not do any work, but should merely define the presentation of results (i.e. the order of tables and perhaps some commentary).
     Among other things, the rmarkdown template currently creates the excel output, which makes this quite brittle (and required the hack-fix for send attachment #44.
     No dependencies in rmarkdown templates, in particular, or at least not without programmatically rendereing the template as part of the check.
2. documentation in the code
   • should be longer and live as close as possible to the code (also used to be refactor ELT + analysis codebase #52), ideally as roxygen2
   • might have to also be shown elsewhere
3. reduce dependencies
4. tests
5. assertions (on inputs)

This thould address, among others:

• fix inconsistency in no openAPC reported dois #67
• count tdms only once per article #66
• count share of has ORCID instead of T/F #64
• surface and protect against all bad DOIs in report #60
• tu_dois appears to include duplicates #55
• refactor ELT + analysis codebase #52
• enforce unique dois #50
• skip test emails in dev #68
• fix warnings, errors and notes #13
• narrow down snapshot testing #51