v0.2.0

• Major code refactoring for readability and maintainability
• Vastly improved function signatures (breaking changes to v0.1.0)
• Added argument checks for all user-facing functions
• Add cov_vars argument for selecting variables to use for for imputations
  but not for covariance matrix
• Complete control over mice() via ... arguments
• Better function returns, e.g. main function now returns an S3 object of
  class mifa which has nice summary() and print() generics
• Much improved documentation and examples
• Added a README.md with installation and usage instructions
• Added a {pkgdown} site at https://teebusch.github.io/mifa/
• Added a NEWS.md file to track changes to the package.
• Remove unused dependencies
• Added continuous integration and unit tests