Feature Request: Enable non-R code exercises

[lukaskawerau]

In normal Rmd documents you can specify SQL-only chunks, based on a database connection via DBI:

```{r}
library(DBI)
mydata = dbConnect(RPostgres::Postgres(), 
               dbname = "some_database",
               host = 'localhost',
               port = 5432,
               user = 'postgres',
               password = 'mysecretpassword')
```

```{sql, connection=mydata}
CREATE TABLE table_name (
  PersonID int,
  LastName varchar(255),
  FirstName varchar(255)
);
```

However, this does not work with learnr. I would like to be able to do the same in an exercise so that students can directly submit SQL statements:

```{sql, connection=mydata, exercise = TRUE}

```

Where students then would have to enter SQL code into the exercise.

[schloerke]

I'm going to leave this open as it's a good idea.

It's going to be a while as the correct infrastructure is not there.

[coatless]

@schloerke let's generalize to all of R Markdown's supported engines?

Outside of SQL, it would be useful to have Python.

[cassws]

Hi, @coatless @schloerke should I assume that Python isn't working with learnr at this time? Is there any workaround with e.g. reticulate? If not, I would be interested in getting involved with adding this functionality as it would be extremely helpful for a work project (slash helpful in general!)

[schloerke]

@zoews

should I assume that Python isn't working with learnr at this time?

Correct

Is there any workaround with e.g. reticulate?

I would like to use knitr's engine to process the information. R Markdown already has support for python.

This would allow for all existing languages that can produce output to work as expected. (Persistent connections, like shiny applications, will not be possible. Only code that evaluates and returns is possible in the current execution model.)

Recipe

evaluate_exercise is where the execution of the user exercise happens.

I'm guessing that (as of writing this comment)

• the knitr::spin call will have to have a different function call when working with different languages.
• the code pertaining to last_value and learnr_output_handler will need to not be executed
• the checker functions (pre and post) might need an extra argument of language.

Debugging will be hard with a temp file being created each time. I recommend changing this line to a local working path folder.

---

@zoews For implementation ideas/discussion/troubleshooting, please open a PR. 😄

[braunsb]

Just leaving my vote: If we had this feature, we would be using learnr all over the organization!

[TonyFly3000]

would love to have following Feature

{sql, connection=mydata, exercise = TRUE}

[JosiahParry]

Just want to throw in my vote for this as well :)

[schloerke]

@nischalshrestha has re-written how exercises are calculated in #390 (just merged to master). Thank you @zoews for getting the conversation started and showing what is possible.

Currently, no UI changes have been merged into master from #310. Working on this in #397

If you could test this locally, that would be great! If you run into any issues, please comment on #397, as I will be making followup adjustments there. Thank you!

Example polyglot tutorial

---
title: "Multi-language exercises"
author: "Zoë Wilkinson Saldaña"
output: learnr::tutorial
runtime: shiny_prerendered
---

```{r setup, include = FALSE}
library(learnr)

# altered from https://datacarpentry.org/R-ecology-lesson/05-r-and-databases.html
if (!file.exists("portal_mammals.sqlite")) {
  download.file(url = "https://ndownloader.figshare.com/files/2292171",
                destfile = "portal_mammals.sqlite", mode = "wb")
}
mammals <- DBI::dbConnect(RSQLite::SQLite(), "portal_mammals.sqlite")
```

### Python or R?

The following two code exercises run language agnostic code.  It will finally evaluate to either 5 or 10.

```{r r-not-python, exercise=TRUE, exercise.lines = 5}
# if running in R, this should return 10
# if running in Python, this should return FALSE and 5
x = 5
x <-10
print(x)
```

```{python python-not-r, exercise=TRUE, exercise.lines = 5}
# if running in R, this should return 10
# if running in Python, this should return FALSE and 5
x = 5
x <-10
print(x)
```

### Bash

Execute exercises in bash

```{bash bash-not-r, exercise = TRUE}
echo "hello world"
date
```

### SQL

Use a SQL connection defined in a setup chunk.

```{sql sql-not-r, exercise = TRUE, connection="mammals"}
SELECT *
FROM `surveys`
LIMIT 10
```

---

Closing as fixed in #390