function can be named fun- or fun_ or fun-my_function_nameexample can be named ex- or ex_ or ex-my_function_nametests can be named test- or test-my_function_namedevelopment can be named dev- or dev_ or dev-my_function_nameadd_flat_template() for lazy developers: add_additional(), add_full(), add_minimal()Yes. Although this is not the real goal, the flat template can be filled and run like any other Rmarkdown file.
However, you need to keep in mind that this will be transformed as a R package, which requires some specific attention.
development chunk at the beginning of your flat template to declare all packages that you will need to explore your code```{r dev}
library(glue)
library(stringi)
```Note that this will not be used anywhere in the package. A development chunk is only there for your code exploration, during development.
Remember that in the package structure, examples and tests code will be run independently. Thus, examples and tests chunk need to have all code required to run independently.
The development chunk that contains the inflate() call need to have eval=FALSE parameter to avoid side effects if you knit the flat file.
library() for the future vignette ?If you use a classical chunk, without any specific name, it will be copied as is in the vignette.
```{r}
library(glue)
library(stringi)
```I created a chunk example but I do not want the example to be run in the function example and I do not want it to be run in the vignette.
eval=FALSE in the chunk options, so that it wont be run in the vignette#' \dontrun{} syntax, with #' before so that examples in the function example will not be run```{r function-myfunction}
myfunction <- function(x) {
x
}
```
```{r example-myfunction, eval=FALSE}
# Will be run in example but not in vignette
myfunction(10)
# Will not be run in example
#' \dontrun{
myfunction(12)
#' }
#' \dontrun{
#' myfunction(12)
#' }
```function chunk as usualTo document datasets, pkgdoc, special R files, you can write them as is in the function chunk.
If {fusen} does not detect the keyword function() or R6Class() in the function chunk code, then the chunk is copied as is in a “R/” file.
```{r function-cars}
#' cars
#'
#' data in 'datasets'.
#'
#' @format A data frame with 50 rows and 2 variables:
#' \describe{
#' \item{ speed }{ numeric }
#' \item{ dist }{ numeric }
#' }
#' @source Ezekiel, M. (1930) Methods of Correlation Analysis. Wiley.
"cars"
```All chunks named dev or development will not be used in the package.
Use them for your exploration that you do not want to keep at the end of the process.
This won’t affect your package.
```{r dev}
# Some code exploration
```You can use inflate(vignette_name = c("Super title" = "01-Super Slug")) for vignette title different from vignette Entry
inflate(vignette_name = c("Super title" = "01-Super Vignette Index Entry"))inflate(vignette_name = NA)If it is the first function of the flat template:
add_flat_template("minimal", flat_name = "my_function_name"). In this case, the template will be pre-filled with your function name: chunk names, function calls.If this is the second function inside an existing template:
add_fusen_chunks("my_function_name")This can be added in the “dev_history.Rmd”, and will replace a place holder in your flat rmd (HERE in this case) with proper fusen chunks for all your future functions to be.
HERE in your flat file# Path of the flat file
path_flat_rmd <- here::here("dev/flat_minimal.Rmd")
# Name of future functions to add - example
fun_nms <- c(
"get_contract_by_country_of",
"get_contract_subsidies_of",
"get_contract_effort_of",
"get_contract_fees_of",
"get_contract_target_countries_of",
"get_contract_all_info_of"
)
# Create content
l_chunks <- purrr::map_chr(fun_nms, fusen:::build_fusen_chunks)
chunks <- paste(l_chunks, collapse = "")
# Add in the flat file
flat_rmd <- readLines(path_flat_rmd)
flat_rmd <- sub("^HERE$", chunks, flat_rmd)
writeLines(flat_rmd, path_flat_rmd)inflate_all()Read the dedicated vignette “inflate-all-your-flat-files”
vignette("inflate-all-your-flat-files", package = "fusen")There are multiple ways of doing it. Choose one of: use the section title structure, use roxygen tags or use chunk parameter.
Use the 3-set of chunks (fun, example, test) twice, under the same title section of the Rmd
# One title for both groups of chunk
The code and tests for the first function
```{r function-fun1}
```
```{r example-fun1}
```
```{r test-fun1}
```
The code and tests for the second function
```{r function-fun2}
```
```{r example-fun2}
```
```{r test-fun2}
```@rdname in function roxygen# Title 1 for function 1
```{r function-fun_rdname1}
#' My fun_rdname1
#'
#' @param x Vector of Numeric values
#' @inheritParams stats::median
#'
#' @rdname same_rdname
#' @return
#' Median of vector x
#' @export
#'
#' @examples
#' my_fun_rdname1(2:20)
my_fun_rdname1 <- function(x, na.rm = TRUE) {
if (!is.numeric(x)) {
stop("x should be numeric")
}
stats::median(x, na.rm = na.rm)
}
```
# Title 2 for function 2
```{r function-fun_rdname2}
#' My fun_rdname2
#'
#' @param x Vector of Numeric values
#' @inheritParams stats::median
#'
#' @rdname same_rdname
#' @return
#' Median of vector x
#' @export
#'
#' @examples
#' my_fun_rdname2(2:20)
my_fun_rdname2 <- function(x, na.rm = TRUE) {
if (!is.numeric(x)) {
stop("x should be numeric")
}
stats::median(x, na.rm = na.rm)
}
```@filename in function roxygen@filename is recognized only by {fusen} as a proper roxygen tag to merge multiple functions in the same “R/” and “tests/” files.
This code line is removed in the resulting “R/” files to avoid any interference with Roxygenize.
# Title 1 for function 1
```{r function-fun_filename1}
#' My fun_filename1
#'
#' @param x Vector of Numeric values
#' @inheritParams stats::median
#'
#' @filename same_filename
#' @return
#' Median of vector x
#' @export
#'
#' @examples
#' my_fun_filename1(2:20)
my_fun_filename1 <- function(x, na.rm = TRUE) {
if (!is.numeric(x)) {
stop("x should be numeric")
}
stats::median(x, na.rm = na.rm)
}
```
# Title 2 for function 2
```{r function-fun_filename2}
#' My fun_filename2
#'
#' @param x Vector of Numeric values
#' @inheritParams stats::median
#'
#' @filename same_filename
#' @return
#' Median of vector x
#' @export
#'
#' @examples
#' my_fun_filename2(2:20)
my_fun_filename2 <- function(x, na.rm = TRUE) {
if (!is.numeric(x)) {
stop("x should be numeric")
}
stats::median(x, na.rm = na.rm)
}
```filename = "the_comon_filename"Add it in the function chunk only
# Title 1 for function 1
```{r function-fun_chunk1, filename = "fun_chunk1"}
#' The code of your function 1
```
```{r example-fun_chunk1}
```
```{r test-fun_chunk1}
```
# Title 2 for function 2
```{r function-fun_chunk2, filename = "fun_chunk1"}
#' The code of your function 2
```
```{r example-fun_chunk2}
```
```{r test-fun_chunk2}
```During checks, the tests are run relative to the “tests/testthat/” directory.
You will need to anticipate the two ways of reading the data:
# The path relative to the "tests/testthat" directory for tests
the_file <- "my_file.csv"
if (!file.exists(the_file)) {
# The path to use during dev in the flat file
the_file <- file.path("tests", "testthat", the_file)
if (!file.exists(the_file)) {
stop(the_file, " does not exist")
}
}
my_file <- read.csv(the_file)inflate()?Yes. You can run and load function chunks only in the currently opened flat file with load_flat_functions().
With long flat file currently in development, and before inflate(), it is sometimes difficult to run all chunks needed after multiple modifications. This can also be useful when you start again your development the day after.
load_flat_functions() is like a load_all() for a flat file, although it does not account for dependencies.
In the console, run:
fusen::load_flat_functions()You can also run function chunks of a specific flat file with:
fusen::load_flat_functions(flat_file = "dev/flat_full.Rmd")Yes you can. As long as what you include in your qmd flat file is good for a R package vignette, you can use the qmd format.
This will not really change the output of anything as the flat file is not meant to be rendered.
The vignette created from this flat file will still be a Rmd file. But why not!?
Hence, you can add a flat file and change its extension to “.qmd” if you like.
Of course ! That is even recommended as it allows to properly separate the business logic (in “dev/” directory) from the UI (in the modules of the “R/” directory).
Start by creating a {golem} project. Then you can start adding your flat files.
We recommend to create modules with names in relation to flat files, so that you can better navigate in your project.
golem::create_golem('my.golem.app')
fusen::add_minimal_flat(flat_name = "page1")
golem::add_module(name = "page1")Although default {golem} “dev/” files already contain the main actions to maintain your package, you can still add the “0-dev_history.Rmd” file recommended with {fusen} using add_dev_history(). Note that some sections will be redundant with the {golem} dev files.
You can draw the structure of your package with `fusen:::draw_package_structure().
Read vignette “Draw a tree of your package files and functions”" for more details.