• Intro
• Installation
• Demo
• Recommended Workflow
  • Writing the Question Templates
  • Generating the Exercises
    • Tip
    • Special Characters on Windows
  • Uploading to MyLearn
    • First Upload
    • Re-uploading with the Same Shortname
  • Importing Exercises
  • Organizing a Random Question on MyLearn
• Issues
• Maintainance

Randomized multiple-select and single-select question generation for the MyLearn teaching platform used by universities. Question templates in the form of the R/exams package are transformed into MyLearn’s XML format. Note that the feedback has to stay simple, see the Issues section. The imported questions can be organized into exams in the Learning Activities.

Dependencies:

• libxml2: found on popular Linux distros and in Rtools by default
• pandoc: gets installed by R during installation
• R: a recent version
• on Windows: Rtools
• popular R packages: installed automatically

The question import functionality might have to be turned on by your MyLearn support team.

install.packages("exams.mylearn")

One creates an R/exams compatible exercise file similar to example_question.Rmd, decides for a name prefix and for the number of generated exams.

library("exams.mylearn")
n <- 10L  # number of variants of the exercise
name <- "learn_exercise"  # prefix of the generated file names
exercise_file <- example_paths()["R_table"]
output <- exams2mylearn(filename = exercise_file,
                        n = n, dir = ".",
                        outfile = "question_table.zip")

The output .zip filepath is saved in output. One can upload that .zip file to Learn directly (after the Learn development team has activated that feature for the given course).

The question templates are the .Rmd files. Find how-to’s in the files returned by example_paths(). Best starting point is example_paths()["mixture"], that contains solutions to most features you might need in an exam question.

The recommended way of debugging the questions is via

filename <- "potential-exercise.Rmd"
exams::exams2html(filename, n = 1L,
                  converter = "pandoc-mathjax",
                  verbose = TRUE)

The generated HTML file opened in the browser closely resembles the looks of the exercise on MyLearn.

Set the current working directory to where the exercise templates are, then run the following.

filenames <- list.files(pattern = ".*.Rmd$")
n <- 25L
for (filename in filenames) {
  exams2mylearn(filename, n = n,
                outfile = "exam-exercises.zip",
                dir = ".", dontask = TRUE)
}

If you’re uploading improved exercises over and over again, I recommend executing the following for loop instead:

filenames <- list.files(pattern = ".*.Rmd$")
n <- 500L
for (filename in filenames) {
  exams2mylearn(filename, n = n,
                outfile = "exam-exercises.zip",
                dir = ".", dontask = TRUE,
                distort.shortname = TRUE,
                verbose = TRUE)
}

This way the shortname will be unique every time and you will have more feedback on where the generation process is.

The .Rmd file is assumed to have UTF-8 encoding. That is important in case the .Rmd file contains special (e.g. German) characters.

In RStudio, one can choose the encoding for file reading under File > Reopen with Encoding, and for file writing under File > Save with Encoding. It is recommended to do both with UTF-8.

In the course find Administrate, then under “Old” learning materials click on organize. Then under XML-file choose the .zip file created in the previous step. Finally, click on Import learning materials.

At the first upload a newly opened window shows a row for each exercise in the .zip file. Those containing plots show a warning, that is normal. By clicking on Ansicht, the uploaded exercise appears.

There is no way out from here, go back to the home page.

In this case a smaller list of the re-uploaded exercises is shown. Choose all of them and click on Submit. This overwrites the old versions.

Yes, uploading and importing are different steps. In the menu of the course’s Learning activities, click on Import > Import of existing learning materials. A table of uploaded exercises appears.

Use the filters to filter out your exercises. The most useful filters are Already imported set to No and the text input where you can use the unique shortnames.

Select all exercises with one click, then Selected items > Import. This brings you to the imported exercises. Select all and then Selected items > Add to clipboard.

Create a new Poolfolder in the Learning activities and open it. Click on Clipboard > Insert content here and then ok. Release all questions in the poolfolder otherwise the random question won’t work.

Now create a Proxy outside of the poolfolder and associate it with the poolfolder you created. This was the last step, congratulations!

The poolfolder contains the pool of questions that the random question will choose from. It samples with replacement from the pool each time someone opens the question. Poolfolders and their content are always hidden from students.

At this point I recommend that you delete the exercises in the imported view. For that: in the menu of the course’s Learning activities, click on Import > Import of existing learning materials. Delete all exercises.

1. MyLearn handles only single select and multiple select questions.
2. “Answerlist” in the Solution block doesn’t work on MyLearn.

Should you find bugs or request features, please use the Issue tracker on Github.