hw09 ready for grading.

[rainerlempert]

I just created the public repository, so you can download the package!

[HScheiber]

Repository/README:

• Your README is sufficiently descriptive, if a little brief. It gets the job done.

• Just a small tip: in R markdown files, to hide an unwanted output you can use {r, eval = FALSE} in the code header.

The R Package: powers

• Installing with devtools::install_github("rainerlempert/Lempert_Rainer_hw09") was successful, thats the first step!

• Using ?powers brought up the package documentation, which is empty. You should have added package documentation using the same process you did for the individual function documentation but instead writing NULL for the "function name".

Function: square

• I checked out the function documentation using ?square, its the same as the one we all made in class together.

• The function seems to work as intended, but without additional features.

• Using this function with a character vector gives a nice descriptive error message, but the error message for other unintended objects (such as a data frame of characters) is still cryptic. It is better to set up your function so that it checks that your input is numeric (or logical) and outputs an error message otherwise. Here you have set it to check explicitly for character inputs and throw an error if found.

Function: cube

• The function documentation is pretty much the same here. Seems fine to me but there are no extra features. Same comments as before apply here.

Function: pow

• You turned this function into an exported function and gave it sufficient documentation.

Function: boxcox

• I think you should have made better documentation here. I had never heard of the boxcox transform before this class, and your function documentation doesn't help me understand what it is for. When should I use a boxcox transform? What are the limitations? etc.

• One of the examples in your documentation for this function boxcox(-5,0) throws an error when actually executed.

• In the function code you use if (lambda!=0), this could simply be replaced by else.

Data: tenvec

• Theres no documentation for this dataset that comes with your package. Data sets that are exported with a package should have their own documentation.

Vignette

• You have to build your Vignettes with devtools::build_vignettes() for them to be accessable to the end user outside of github.

• Provides examples of all your functions, though it's pretty short.

Overall: You have completed most of the tasks required for this assignment, but there are missing parts.

Cheers,

Hayden

[bcahn7]

Hello, @rainerlempert . Here are some comments on your hw09. Hope it helps!

• Your README and vignettes have all the required information to understand the package. More explanations would help users understand the package and the functions.

• I don't know why but ?powers, ?square, ?cube, ?pow, ?boxcox did not work for me.

• It would be a little bit better if you could describe more about the package in powers-package.r.

• You have an internal function, pow(). When functions are more complicated and have some repeating works, internal functions would be more useful to avoid redundancy.

• square(), cube(), boxcox()

  • All the functions are well documented. All of them are informative.
  • All the functions have if statements for a validity check of function input. They work nicely! When the input is a string, the functions stop and warn that these functions require a vector or number.
  • There is a validity check in boxcox() to get a positive input.
  • Your functions have all the essential arguments but it would be better if you could add some more arguments such as plot_it or na.omit.

• Unit tests for every function might check missing parts and reduce warnings or errors.

• It would be good if you could try to import some other packages such as ggplot2 or stringr.

Overall, I think you understand the essential things to make your own package! It would be better if you could experiment a little bit more. You did a good job!