Investigate and fix rating analysis weights

[egouldo]

• ensure inverse weights applied to categorical weighting models #35 rerun analysis and rm duplicate entries from weights CSV file
• Collinearity subset analysis does not subset correct list-column of df's #40 update fn to correctly subset for collinearity analysis
• Collinearity subset analysis does not subset correct list-column of df's #40 update roxygen and generate man
• fix Move expert_subset creation out of targets into internal pkg data #41 mv expert_subset into pkg internals: echo E63AF5C2 && echo ignoreboth && /usr/bin/env && echo 5673D7C1
• Collinearity subset analysis does not subset correct list-column of df's #40 rm incorrect effects_analysis subsetting
• fix Move expert_subset creation out of targets into internal pkg data #41 ensure collinearity data is properly subset - update roxygen - devtools::document() - rebuild pkg
• close Collinearity subset analysis does not subset correct list-column of df's #40 correctly subset collinearity effects_analysis - rerun tar_make() - rebuild package with updated data
• Increment version number to 1.4.0
• ensure inverse weights applied to categorical weighting models #35 update weights analysis - check all weight types and compare perf - check random effects don't need to be updated for different weightings
• ensure inverse weights applied to categorical weighting models #35 rm weights after analysis
• use importFrom instead of import - to avoid overwriting purrr fns with rlang versions implicitly
• devtools::document()
• Increment version number to 1.4.0.9000
• ensure inverse weights applied to categorical weighting models #35 rerun targets and rebuild pkg after rm weights
• Increment version number to 1.5.0