callr | Call R from R | Reactive Programming library

This question is in the continuity of this one: Is it possible to stop executing of R code inside shiny (without stopping the shiny process)?.

The plot that I display in my app takes some time to produce, and I want the users to be able to stop its creation (for instance if they made a mistake in the options). I found this blog post about using callr in Shiny. The workflow is the following:

• create an empty list of jobs/plots
• clicking on "start" creates a background process to create the plot
  • if the user doesn't do anything, the plot is computed in the background. I use invalidateLater() every second to check if the background process is finished. If it is, then I display the plot.
  • if the user clicks on "stop" before the end of the process, the process is killed, removed from the list, and the previous plot is displayed (if there was no plot produced before, nothing is displayed)

First, I'm not sure how this would scale when several people use the app at the same time. Since every background process is independent, I don't think one user would be blocking the others, but I may be wrong.

Second, I'd like to show a waiting indicator on the plot. So far, I used the package waiter to do that, but the problem here is that renderPlot() is being invalidated every second to check if the background process is finished. Therefore, waiter appears and disappears repeatedly as the output is being invalidated.

Below is an example app that mimics the behavior I'd like to have:

The scenario I'm emulating with the below minimal example is allowing a user to engage with a Shiny App (click the numericInput control and see server-side events occur) while a long-running download is occurring (simulated with Sys.sleep(10) within downloadHandler).

In a synchronous setting, when the "Download" button is clicked, the user can still interact with UI elements, but other Shiny calculations (in this case, renderText), get put in a queue. I'd like the asynchronous setting, where the download occurs in the background, and users can still interact with the UI elements and get desired output (e.g. renderText).

I'm using callr::r_bg() to achieve asynchronicity within Shiny, but the issue is that my current code of the downloadHandler is incorrect (mtcars should be getting downloaded, but the code is unable to complete the download, 404 error message), I believe it's due to the specific way in which downloadHandler expects the content() function to be written, and the way I've written callr::r_bg() is not playing nicely with that. Any insights would be appreciated!

Reference:

https://www.r-bloggers.com/2020/04/asynchronous-background-execution-in-shiny-using-callr/

Minimal Example:

BLUF: I am struggling to understand out how to use batching in the R targets package to improve performance in a static and dynamic branching pipeline processed in parallel using tar_make_future(). I presume that I need to batch within each dynamic branch but I am unsure how to go about doing that.

Here's a reprex that uses dynamic branching nested inside static branching, similar to what my actual pipeline is doing. It first branches statically for each value in all_types, and then dynamically branches within each category. This code produces 1,000 branches and 1,010 targets total. In the actual workflow I obviously don't use replicate, and the dynamic branches vary in number depending on the type value.

I'm trying to set up a workflow that involves downloading a zip file, extracting its contents, and applying a function to each one of its files.

There are a few issues I'm running into:

1. How do I set up an empty file system reproducibly? Namely, I'm hoping to be able to create a system of empty directories to which files will later be downloaded to. Ideally, I'd like to do something like tar_target(my_dir, fs::dir_create("data"), format = "file"), but I know from the documentation that empty directories are not able to be used with format = "file". I know I could just do a dir_create at every instance which I need it, but this seems clumsy.

2. In the reprex below I'd like to operate individually on each file using pattern = map(x). As the error suggests, I'd need to specify a pattern for the parent target, since format = "file". You can see that if I did specify a pattern for the parent target, I would again need to do it for its parent target. As far as I know, a pattern cannot be set for a target that has no parents (but I have been wrong many times before).

I have a feeling I'm going about this all wrong - thank you for your time.

Since a few months, ggplot2 started to save png files with a transparent background. The code output in Rstudio and when saved as pdf looks great. It happens mainly with the use of themes when I omit the gray panel background. I tested it on my macbook with "Preview" and on a Windows Computer with the "foto viewer" there.

I was wondering what happens when callr_function = NULL? Is it just issues with things maybe being in the environment/side effects?

Mainly wondering because I was passing quite large spatio-temporal arrays (0.5 to 5 gigs) and callr serialization via saveRDS is quite slow.

The two things I was thinking about was forking callr and dropping in a different save function or just using callr_function = NULL.

I’m creating a Shiny app that uses the caret package to do some SVM free-text analysis.

The app runs fine without any error in my computer. I’m using R x64 4.0.4 and R studio 1.3.1093

I’m deploying app to an internal enterprise server https://rconnect.xxxx.com/connect/#/apps/####

This app is deployed in the server and started.
But when I reach the line where I run the train function:

The example below is a simple one which tries to assert the column y is always positive (y>0). How can I extract the errored data (row 3 with the negative value,into a dataframe maybe, or any convenient object) while allowing the workflow to continue with "cleaned" data?

I ran into a missing graph issue while developing histograms and geometry bar plots in r using the ggplot function. The issue is intermittent and occurs across multiple data sets. One data set, the largest, is described here.

My initial code, used to generate randomized data, is this: