stevenlshafer/stanpumpr

stanpumpR, and PK/PD simulation program

stanpumpr's Issues

Oxycodone IV model

As discussed at WFSA session

Available in Australia and New Zealand

Button colours should be more intuitive

I think there should be a more consistent colour scheme, where "Cancel" buttons are just grey and "Delete" buttons are red.

For example, double click on the plot to add a drug, the Cancel button is orange and delete is green. Those colours make give the user too many things to focus on and it's distracting.

Double click to add drug: only show applicable units

When double clicking on the plot to add a drug, all the units are shown instead of only the applicable units

Effect and plasma site targeting mode

Graph bolus and infusion rate over time to achieve a target (reverse of current mode)

May be in useful in research for situations where targets are recorded, not infusion rates.

sendSlide() leaves behind PowerPoint, Excel and PNG files after sending the slides

These temporary files should be unlinked after sending the email.

Reference time is not persisted through bookmarks

Although it appears that the list of reference times options is static, it is not as it can be overriding either manually or automatically with the current time. This is could be fixed by altering onRestored() (complex) or the field could be replaced with a textInput() (simple).

MIT license contradicted by README

The package is under MIT license, which explicitly allows commercial use.

Yet the following excerpt from the README attempts to restrict commercial use and even states "all rights reserved".

Companies seeking to develop such systems should contact Dr. Shafer to request written permission to incorporate stanpumpR into their products. 
Without written permission, stanpumpR must not be incorporated into commercially available systems.

It is inconsistent and, although I am not a lawyer, I would imagine the MIT license superseeds your note anyway.

Add multiple models for each drug

The ideal implementation would be able to demonstrate multiple plasma and effect site concentration prediction lines, with one for each model.

The tool could then be used to compare the outcomes of various models, which would be excellent for teaching.

Input covariates should be range and type checked to prevent crashes when invalid data is input

Crash on selecting "Log Y axis"

There is an immediate crash on choosing the "Log Y axis" option with any input data.

Add support for shinylive

As the documentation reads: "Shinylive allows you to run Shiny applications entirely in a web browser, without the need for a separate server"

stanpumpR should be faster, more responsive and more reliable when run directly in the browser. As it has no external datasource dependencies, this is the ideal way to run it!

https://github.com/posit-dev/r-shinylive

Modal intro text shows on server reconnect

This is very disruptive when using application for an extended period and the server connection is interrupted (for example, by a shinyapps.io server sleeping, or a bad internet connection)

Email slides won't send

Reported by @StevenLShafer.

stanpumpR is no longer able to send e-mail. This is the error message from the stanpumpR log:

2022-10-17T20:07:03.758416+00:00 shinyapps[861505]: Address is OK
2022-10-17T20:07:07.425484+00:00 shinyapps[861505]: input$sendSlide 1
2022-10-17T20:07:07.428221+00:00 shinyapps[861505]: Sending email to [email protected]
2022-10-17T20:07:07.429311+00:00 shinyapps[861505]: In function sendSlide()
2022-10-17T20:07:08.521821+00:00 shinyapps[861505]: Saving PPTX
2022-10-17T20:07:08.544217+00:00 shinyapps[861505]: Starting ggexport()
2022-10-17T20:07:08.549076+00:00 shinyapps[861505]: [1] "Slides/Preview.1.221017-200707%!d(MISSING).png"
2022-10-17T20:07:09.144123+00:00 shinyapps[861505]: file saved to Slides/Preview.1.221017-200707%!d(MISSING).png
2022-10-17T20:07:09.144521+00:00 shinyapps[861505]: Fixing Units for export
2022-10-17T20:07:09.145349+00:00 shinyapps[861505]: Creating workbook
2022-10-17T20:07:09.156248+00:00 shinyapps[861505]: Writing covariates
2022-10-17T20:07:09.189637+00:00 shinyapps[861505]: Writing dose table
2022-10-17T20:07:09.446145+00:00 shinyapps[861505]: Writing simulation results
2022-10-17T20:07:09.453801+00:00 shinyapps[861505]: Writing results for plotting
2022-10-17T20:07:09.461151+00:00 shinyapps[861505]: Writing PK parameters

2022-10-17T20:07:09.461682+00:00 shinyapps[861505]: Drug =
2022-10-17T20:07:09.464333+00:00 shinyapps[861505]: Error in value[3L] :
2022-10-17T20:07:09.464343+00:00 shinyapps[861505]: (converted from warning) Error in which: object 'drugDefaults_global' not found
2022-10-17T20:07:09.464348+00:00 shinyapps[861505]: Calls: local ... tryCatch -> tryCatchList -> tryCatchOne ->
2022-10-17T20:07:09.464695+00:00 shinyapps[861505]: Error detected in stanpumpR
2022-10-17T20:07:09.464705+00:00 shinyapps[861505]: Error in value[3L] :
2022-10-17T20:07:09.464717+00:00 shinyapps[861505]: (converted from warning) Error in which: object 'drugDefaults_global' not found
2022-10-17T20:07:09.464730+00:00 shinyapps[861505]: Calls: local ... tryCatch -> tryCatchList -> tryCatchOne ->

Incorrect conversion from Inches to centimeters

Hello, it looks like the wrong unit conversion between inches and cm, shouldn't it be 2.54, not 2.56?

stanpumpR/R/globalVariables.R

Lines 12 to 13 in ffc01a9

	UNIT_CM <- 1
	UNIT_INCH <- 2.56

Crash on clicking plot when "Time until _" is ticked

You can no longer update the emergence concentration or get the emergence time line to appear. It is unclear if this is related to recent changes in stanpumpR or Shiny. This is possibly related to #66 ?

purr 1.0.0 breaks the simulation plot

purr 1.0.0 breaks the simulation plot

"Automatic coercion from double to character was deprecated in purrr 1.0.0.
ℹ Please use an explicit call to as.character() within map_chr() instead."

https://www.tidyverse.org/blog/2022/12/purrr-1-0-0/

Crash on load when running on shinyapps.io

Bug introduced by PR #80.

The '%btwn%' function is actually used and should not have been deleted. Deleting this function causes a crash when the application is run on shinyapps.io.

"Set ventilation/emergence concentration" is not shown when adding a drug by double clicking the plot

I don’t know what the “Set ventilation/emergence concentration” input does, but I noticed that it shows up when you single click on the plot to add a drug but it doesn’t show up when you double click on the plot

Rename master branch

Rename master branch to main given guidance on https://github.com/github/renaming.

Can't deploy to production due to name conflict

GitHub action workflow needs forceUpdate = TRUE

Covariates that aren't implemented should be disabled in standard app view to avoid confusion

Shiny error from ggplot 3.4.0 deprecating size attribute warning

App fails to load due to new warning being treated as an error

Plasma concentration not appearing

Move all instances of 'cat()' debugging statements to 'outputComments'

There are many cat() statements in the code. @StevenLShafer is there a reason why some are using cat and some use outputComments?

outputComments is nice because we can easily turn them off and turn on specific ones especially now with the envvar

Bug with "Max Time" input

To reproduce:

• Have only one drug in the table
• Set the time to 20
• Set "Max Time" to 10 (it was previously 60 by default)
• You'll get a nothing plot. Now set Max Time back to 60 -- nothing happens. You must set Max Time to some other value and then back to 60 in order for the plot to come back

Reference time input not adding ":" automatically

Type (for example) "1030" into the reference time input and it will not automatically correct it to "10:30" like the other time input fields. The graph therefore has the wrong labels, "NA:NA".

drugs reactive variable is set from the output of two different functions

The reactive value drugs is set from both of these functions. There should be enforcement that these two functions return the same data types.

stanpumpR/server.R

Lines 248 to 256 in 9fe8753

	newDrugs <- recalculatePK(
	isolate(drugs()),
	drugDefaults(),
	age = age(),
	weight = weight(),
	height = height(),
	sex = sex()
	)
	drugs(newDrugs)

stanpumpR/server.R

Lines 392 to 400 in 9fe8753

	newDrugs <- processdoseTable(
	DT,
	ET,
	drugs(),
	plotMaximum,
	prior,
	plotRecovery
	)
	drugs(newDrugs)

Remove broken Travis CI integration

Travis CI has been replaced with GitHub Actions for CI testing

Setting the error option is itself causing an error because of not being in a reactive context

stanpumpR/server.R

Lines 14 to 23 in 9fe8753

	options(error = function() {
	x <- geterrmessage()
	if (!isShinyLocal) {
	cat("Error detected in stanpumpR\n")
	cat(x, "\n")
	cat("URL: ", prior$url, "\n")
	sendError(url = prior$url, errorMessage = x)
	}
	options(error = NULL)
	})

This causes an error because it references a reactive value not in a reactive context

Test sendMail