debugging R

class: center, middle

![:scale 30%](/assets/images/coding_club_logo_1.png)

<!-- Adjust the presentation to the session. Focus on the challenges,
 this is not a coding tutorial.

Note, to include figures, store the image in the `/docs/assets/images/yyyymmdd/`
    folder and use the jekyll base.url reference as done in this template
    or see https://jekyllrb.com/docs/liquid/tags/#links.
    using the scale attribute ![:scale 30%](...), you can adjust the image size.
-->

# 25 AUGUST 2020

## INBO coding club

Somewhere in the Milky Way

---
class: center, middle

![:scale 90%](/assets/images/20200825/20200825_debuggingR.png)

---
class: center, middle

![:scale 100%](/assets/images/20200825/20200825_debugging_phdcomics.gif)

---
class: center,middle

From the fortunes of the R community: 
![:scale 100%](/assets/images/20200825/20200825_intro_debugging.png)

---
class: left, middle

You have some code, it doesn't work as expected...
<center>![:scale 100%](/assets/images/20200825/20200825_print_in_action1.png)</center>
 You start to search by inserting `print()` instructions here and there..

---
class: left, middle

Welcome to the `print()` FESTIVAL!

<center>![:scale 100%](/assets/images/20200825/20200825_print_in_action2.png)</center>
 
And what if the error is not in your code but in a function embedded in it?

---
class: center, middle

![:scale 90%](/assets/images/20200825/20200825_RStudio_IDE_cheatsheet.png)
[Download cheatsheet here](https://github.com/inbo/coding-club/blob/master/cheat_sheets/20180821_cheat_sheet_rstudio-ide.pdf)

---
class: center, middle

![:scale 90%](/assets/images/20200825/20200825_debug_mode_zoom_cheatsheet.png)

---
class: center, middle

![:scale 90%](/assets/images/20200825/20200825_debug_mode_zoom_cheatsheet2.png)

---
class: left, middle

## Why debugging?

- save time: "debugging" using `print()` is time consumming  + you have to clean up your code after
- jump in **loops**, in **if-else** statements, in **functions**

---
class: center, middle

### How to get started?

Check the [Each session setup](https://inbo.github.io/coding-club/gettingstarted.html#each-session-setup) to get started.

### First time coding club?

Check the [First time setup](https://inbo.github.io/coding-club/gettingstarted.html#first-time-setup) section to setup.

---
class: left, middle

![:scale 100%](/assets/images/coding_club_sticky_concept.png)
 No yellow sticky notes online :-( We use hackmd (see next slide) but basic principle doesn't change.

---
class: center, middle

### Share your code during the coding session!

Go to https://hackmd.io/Nb8AFC5gQ7GMJ4FyiTYqow?both

---
class: left, middle

# Download data and code

Today we will use a R script. Download* it in your folder `src/20200825`.

- R script: [20200825/20200825_challenges.R](https://github.com/inbo/coding-club/blob/master/src/20200825/20200825_challenge.R)

This time we don't have any dataset to download.

* __Note__: check the getting started instructions on [how to download a single file](https://inbo.github.io/coding-club/gettingstarted.html#each-session-setup)

---
class: left, middle

# A live code intro

![:scale 83%](/assets/images/20200825/20200825_livecoding_snippet.png) 
```r
Error in FUN(newX[, i], ...) : is.atomic(x) is not TRUE
```

---
class: left, middle

# A memory aid

## Traceback your code

`traceback` your code to find where it stops to work.

In RStudio: in **Debug pane** -> **On Error** -> **Error Inspector** and then click on **Show Traceback** in the error message 
![:scale 100%](/assets/images/20200825/20200825_livecoding_show_traceback.png)
In basic R:
```r
option(error = traceback) # trace the error back, better than option(error = message)
```

---
class: left, middle

# A memory aid

## Debug your code: activate debugging mode

### 1st alternative

Source the file and set breakpoints (red circles) where you think something wrong happens

![:scale 50%](/assets/images/20200825/20200825_livecoding_show_breakpoint.png)

---
class: left, middle

# A memory aid

## Debug your code: activate debugging mode

### 2nd alternative

Set `options(error = recover)` and select the function:

![:scale 100%](/assets/images/20200825/20200825_livecoding_show_error_recovery.png)

---
class: left, middle

# A memory aid

## Debug your code: activate debugging mode

### 3rd alternative

Activate debugging mode on a specific function by using `debug(function_name)` or `debugonce(function_name)`.

```r
debug(calc_var) # or debugonce(calc_var)
```

Once this function is debugged, run `undebug(function_name)`.

---
class: left, middle

# A memory aid

## Debug your code: work in debugging mode

Once you are in debugging mode, use **Next** (n) to move to next line, **Continue** (c) to go to next breakpoint (red circles)

![:scale 70%](/assets/images/20200825/20200825_livecoding_show_debugging_mode_actions.png)

---
class: left, middle

# 123 magic

Write down a positive integer on a piece of paper:

```
886328712442992
```

Count up the number of even and odd digits, and the total number of digits:

```
10 5 15
```

String the digits of those three numbers together to make a new number:

```
10515
```

Perform the same operation on the obtained number:

```
1 4 5 ⟶ 145
```

And keep iterating:

```
1 2 3 ⟶ 123
```

You'll always arrive at the number 123!

Credits: 
(EN) https://dodona.ugent.be/en/courses/239/series/2168/activities/192047393/ 
(NL) https://dodona.ugent.be/nl/courses/239/series/2168/activities/192047393/

---
background-image: url(/assets/images/background_challenge_1.png)
class: left, middle

# Challenge 1

Use debug techniques to find (and possibly solve) the bug in the function `evenOdd()` in [20200825_challenge.R]():

```r
#' Function evenOdd() takes an integer n and returns a list containing two integers
#' that respectively indicate how many even and odd digits occur in n
#'
#' @examples
#' evenOdd(886328712442992) -> (10, 5)
#' evenOdd(10515) -> (1, 4)
#' evenOdd(145) -> (1, 2)
evenOdd <- function(n) {
 char_n <- as.character(n)
 counter_even <- 0
 counter_odd <- 0
 for (i in char_n) {
 digit <- as.integer(i)
 if (digit %% 2 == 0) {
 counter_even <- counter_even + 1
 } else {
 counter_odd <- counter_odd + 1
 }
 }
 return(list(n_even = counter_even,
 n_odd = counter_odd))
}
```

---
background-image: url(/assets/images/background_challenge_2.png)
class: left, middle

# Challenge 2

Use debug techniques to find (and solve) bugs in functions `step()` and `steps()` in [20200825_challenge.R]()

```r
#' Function step() takes a positive integer:
#' 886328712442992
#' 1. Count up the number of even and odd digits, and the total number of digits:
#' 10 5 15
#' 2. String the digits of those three numbers together to make a new number:
#' 10515
#' 3. Return it as a number
step <- function(n) {
 n_odd_even <- evenOdd(n)
 total_digits <- nchar(n)
 n_odd_even_total <- n_odd_even
 n_odd_even_total$n_total <- total_digits
 return(n_odd_even_total)
}

#'Function `steps` takes an integer and return how
#' many steps you need before the number 123 is reached.
steps <- function(n) {
 while (n != 123) {
 n <- step(n)
 n_steps <- n_steps + 1
 }
 return(n_steps)
}
```

---
class: left, middle

# Intermezzo: defensive programming

Debugging starts already while writing code! Defensive programming is the best way to solve a good bunch of (future) bugs.

1. Add pre-conditions to your functions to avoid unexpected and unpredictable errors in later stages (**fail fast principle**: stop the current operation as soon as any unexpected error occurs)
```r
# PSEUDOCODE !!
if (date < today) stop("Date is too early");
if (class input != numeric vector) stop("Invalid type of input: only numeric vectors allowed.")
```

2. Assert statements to validate output at intermediate steps, e.g. output of 3rd party libraries that you use in your project.

---
background-image: url(/assets/images/background_challenge_3.png)
class: left, middle

# Challenge 3

Apply defensive programming to avoid unexpected behavior or cryptic error messages ("fail fast" principle).

Function crashes with:

1.  "very large" numbers
![](/assets/images/20200825/20200825_challenge3_unexpected_error_big_numbers.png)
2. decimal numbers
![](/assets/images/20200825/20200825_challenge3_unexpected_error_decimal_number.png)
3. vectors or list of numbers
![](/assets/images/20200825/20200825_challenge3_unexpected_error_multiple_numbers.png)

---
class: left, middle

# Bonus challenge

It's not the first time we speak about debugging. We did it already in 21 Aug 2018, so there are other three challenges waiting to be solved! See [slides](https://inbo.github.io/coding-club/sessions/20180821_coding_club_debugging.pdf) (pdf). Download [data](https://github.com/inbo/coding-club/tree/master/data/20180821) and [R scripts](https://github.com/inbo/coding-club/tree/master/src/20180821) as links in slides are broken after we moved from Google Drive to GitHub.

* __Note__: check the getting started instructions on [how to download a single file](https://inbo.github.io/coding-club/gettingstarted.html#each-session-setup)

---
class: left, middle

# Miscellaneous tips

1. Smaller functions are better: easier to debug, easier to understand
1. Find and avoid **global variables**
```r
library(codetools)
f <- function(z) {y <- 3; print(x + y + z)} # BAD FUNCTION
findGlobals(f)
## [1] "<-" "{" "+" "print" "x"
```
1. Don't hard code numbers: use variables instead (n of iterations, parameter values in simulations, ...). Easier to read and reduce bugs when you use the same number multiple times
1. Build up code in pieces, testing along the way
1. Make big changes in small steps

---
class: left, middle

## Resources

- [webinar](https://www.youtube.com/watch?v=-yy_3htRHdU&t=444s) and related [tutorial](https://htmlpreview.github.io/?https://github.com/berkeley-scf/tutorial-R-debugging/blob/master/R-debugging.html) from Chris Paciorek, Department of Statistics, UC Berkley
- [Debugging in RStudio tutorial](https://support.rstudio.com/hc/en-us/articles/205612627-Debugging-with-RStudio) written by RStudio itself
- [Debugging chapter](https://adv-r.hadley.nz/debugging.html) from Hadley Wickham's book Advanced R, 2nd edition
- [123 exercise](https://dodona.ugent.be/en/courses/239/series/2168/activities/192047393/) from Dodona, the programming platform hosted by UGent
- Video recording of the coding club session is available on [INBO's vimeo channel](https://vimeo.com/456879450)

---
class: center, middle

![:scale 30%](/assets/images/coding_club_logo_1.png)

Room: Herman Teirlinck - 01.72 - Kaat Tilley (?) 
Date: __24/09/2020__, van 10:00 tot 12:00 
Subject: be proficient with stRings
(registration announced via DG_useR@inbo.be)