lukesRlib v0.2.12

Bug fixes

• get_assoc() - Drop empty factor levels if using haven::as_factor()
• get_assoc() - Correctly identify covariate variable names when using haven::as_factor()

Changes

• get_assoc() - New option "include_formula" allows user to request that the full regression formula (with covariates) is included in the output
• get_assoc() - Outputs exposure/outcome/covariate variables not in data.frame when verbose=TRUE (shows first 6 when FALSE)
• get_assoc() - Prints a bit more output to the screen by default

lukesRlib v0.2.11

Changes

• tidy_ci() now recognises and tidies factor naming if haven::as_factor() is used instead of the base R as.factor()
• get_assoc() now uses haven::as_factor() instead of base R as.factor() for categorical exposures so that labels are shown nicely
• Update dependencies to include {haven} and be more explicit which tidyverse packages are desired
• Update example data categorical variables to include labels
• Fix NEWS formatting

lukesRlib v0.2.10

Changes

• Change URL to reflect my GitHub username change from lukepilling to lcpilling to be more consistent between different logins, websites, and social media
  -- https://lcpilling.github.io/lukesRlib
  -- https://github.com/lcpilling/lukesRlib
• get_assoc()
  -- Add options to Winsorize the exposure or outcome
  -- Remove beep
  -- Improved verbose reporting

lukesRlib v0.2.9

Bug fixes

• get_assoc() - Fix fit_stat_se if missing
• get_assoc() - Fix covariate checking if interaction included in formula. Resolving issue #1

lukesRlib v0.2.8

• get_assoc()
  • add interacts_with argument. Default is "". User can provide a variable name (string) to also get interaction terms
  • Fixed bug in return_all_terms when exposure was factor
  • Include SE for C-statistic if returning fit_stat for coxph() models

lukesRlib v0.2.7

• get_assoc() - add return_all_terms argument. Default is false. If TRUE, adds a new col terms and returns the estimates for all independent variables in the models.
• get_assoc() - remove subset_d argument
• get_assoc() - fixed if exposure or outcome needs to be protected by backticks (previously failed if variable was something like "bmi-1")
• get_assoc() - fixed problem where covariate estimates were also included in certain situations (e.g., if x="age" do not get covariate "percentage")
• Update examples for many functions

lukesRlib v0.2.6

Since last release:

• get_assoc() - fixed if no covariates provided
• get_assoc() - if x and y are the same, skip (useful if providing same list of variables twice for pairwise analysis)
• get_assoc() - added verbose option for bug checking
• get_assoc() - added beep option for beeping when finished
• tidy_ci() - add some functionality for summarizing fixed-effects from lme4::lmer()

lukesRlib v0.2.5

• get_assoc() now expects the data as first argument, if not defined. Brings it more into the tidyverse. For instance, one can now:
  • data |> filter(age>60) |> get_assoc(x="BMI", y="diabetes", z="sex")
• Add theme theme_minimal_modified for ggplots
• Moves most dependencies to "Imports" - this is the preferred way to use existing packages https://r-pkgs.org/dependencies-mindset-background.html#sec-dependencies-namespace
• Fix -log10p calculation

lukesRlib v0.2.3

• Decided to remove genetics-related functions to instead have them in a separate package, as the dependencies were getting out of hand
  • See https://github.com/lukepilling/gwasRtools
• Added annotate_textp() function for use in ggplot2 plots for adding easy text annotation geoms
• Moved some functions to be internal only to tidy up the namespace
• Fix get_p_extreme if NaN is passed rather than NA

lukesRlib v0.2.2

• Decided to remove doCoxSplinePlot() because it is so much easier to do in {ggplot2} that a whole function is no longer necessary
• • Therefore the {pspline} dependency can be removed
• Set extreme_ps default to FALSE in get_assoc()
• Added p-value threshold option to get_loci()
• Fix problem getting n in get_assoc() when model="coxph"