Changelog

VisXplore 2.1.0

Shiny app improvements

• Variable checkboxes now filter the network plot, correlation matrix, and statistics tabs reactively
• Significance radio buttons now work correctly (fixed string-to-numeric conversion)
• Statistics tab handles both all-numeric (VIF) and mixed-type (GVIF) data
• Added variable checkbox panel to the Statistics tab for interactive feature selection
• Fixed crash on launch caused by lazy evaluation of the data argument
• Clear error message when calling VisXplore() without data
• Fixed xtfrm / “undefined columns selected” errors on the Categorical variables tab
• Suppressed stat_bin() warning by setting explicit bin count in histograms
• Fixed duplicate HTML element IDs between uiOutput containers and their inputs

Code recipe tab

• New “Code” tab in the Shiny app accumulates a reproducible R pipeline as users apply transformations
• Recipe emits standard dplyr::mutate() calls (not package-specific wrappers), so the output is portable
• Includes a “Copy to clipboard” button

API simplification

• Removed exported transform helpers: visx_transform(), visx_ratio(), visx_mean_vars(), visx_collapse_levels()
• Users should use dplyr::mutate() for preprocessing before passing data to pairwise_cor()
• The Shiny app’s Code tab generates the equivalent dplyr code automatically

plot.visx_cor fix

• plot() on a visx_cor object now displays the plot (previously returned invisibly without rendering)

Testing

• Added testServer() tests for Shiny reactive logic (transformations, code recipe, variable filtering, significance)
• Added shinytest2 integration smoke test
• Added shinytest2 to Suggests

VisXplore 2.0.0

Package renamed: VisX → VisXplore

• Package renamed to VisXplore and moved to CU-Anschutz-GMWG GitHub organization
• Main function renamed VisX() → VisXplore(); VisX() still works but is deprecated
• Forked from yingljin/VisX with full git history preserved

S3 class system

• pairwise_cor() now returns a visx_cor S3 object with print(), summary(), plot(), and as.data.frame() methods
• Users can now do result <- pairwise_cor(df); plot(result); summary(result) without launching Shiny
• data_check() now returns a visx_check S3 object with print() method

New exported functions

• get_r2(): R-squared / pseudo-R-squared per variable
• data_check(): data validation utility
• visx_transform(): apply log/sqrt transformations to columns
• visx_ratio(): compute ratio of two columns
• visx_mean_vars(): row-wise mean of selected columns
• visx_collapse_levels(): collapse factor levels

Bug fixes

• Fixed guard condition in pairwise_cor() that always evaluated to TRUE
• Fixed redundant computation of symmetric pairs in pairwise_cor()
• Handle single-numeric-variable case in pairwise_cor()

Code improvements

• Replaced deprecated mutate_all()/mutate_at() with across()
• Vectorized edge-building loops in npc_mixed_cor()
• Expanded test suite with edge cases and new test files