Behind the scenes of the interactive Tauc plot evaluator

The Tauc fitting algorithm in uvvistauc::tauc() can be fragile and return errors if the input parameters don't make sense for the UV-Vis spectrum in question.

In the first iteration of this reactive app (which immediately called tauc() with whatever parameters were set in the app), this frequently led to the user being met by an irrelevant error message in red text instead of a plot.

I considered using shinyglide to achieve the same effect, but I think hiding/showing (or otherwise drawing over) the plot is simpler and I'm happy that this approach worked out.

With the current implementation, observeEvent(input$file1, {...}) combined with observeEvent(input$buttonTauc, {...}), when uploading new data in the same session the previous Tauc plot re-appears briefly before being replaced by the correct one. I can live with this, but ideally we should address this issue.

Links and notes

• https://stackoverflow.com/questions/23233497/outputting-multiple-lines-of-text-with-rendertext-in-r-shiny
• https://shiny.rstudio.com/reference/shiny/latest/textoutput
• https://shiny.rstudio.com/reference/shiny/latest/fileinput
• https://gist.github.com/nutterb/7d723f64201c9afa8bb57774b47bfd7c
• https://stackoverflow.com/questions/20333399/are-there-global-variables-in-r-shiny
• https://stackoverflow.com/questions/30294928/r-shiny-how-to-share-variables-between-rendering-functions
• https://bookdown.org/yihui/rmarkdown/
• https://plotly-r.com/overview.html
• https://shiny.rstudio.com/tutorial/