get test_check running w/o error

Tests should:
- Test whether the function behavior matches the documentation
- Cover all probable use cases for a function or file
- Avoid user/developer interaction (this will require changes from some development-phase tests)
- Avoid producing plots or text output other than testthat output (this will require changes from some development-phase tests)
- Contain internal comments where necessary to explain/justify tests
- Be organized into readable test_that blocks
- Run reasonably quickly
- Use only data available within the package and/or reproducibly generated (e.g., through fixed text definitions of the data, or using set.seed() before random data generation)
- Run cleanly on package tests (Ctrl-Shift-T) and eventually also check (Ctrl-Shift-E)

To do:
- [x] 01-match.arg.loadflex (#26)
- [x] 02-text (#27)
- [x] 03-unit.conversions (#28)
- [x] 04-uncertainty (#29)
- [ ] 05-diagnostics (#30)
- [ ] 06-covariance (#31)
- [ ] 11-metadata (#32)
- [ ] 21-loadModel (#33)
- [ ] 22-loadLm (#34)
- [ ] 23-1-interpolations (#35)
- [ ] 23-2-loadInterp (#36)
- [ ] 24-1-loadReg (#37)
- [ ] 24-2-loadReg2 (#38)
- [ ] 25-loadComp (#39)
- [ ] 31-loadModelInterface (#40)
- [ ] 32-getResiduals (#41)
- [ ] 33-aggregateSolute (#42)
- [ ] 41-predictors (#43)
- [ ] 42-outliers (#44)
- [x] 43-leverage (#45)
- [ ] 44-plots (#46)