In progress
In this tutorial, we will go through ways to
There are many fun ways to do it, here I am more focused for R developers.
Before you create any, make sure you don’t re-invent the wheel and use the best base image for your container as your tool chain may save your lots of time later on.
Official R Docker images is called “Rocker”" project and is on github, please visit the page to find more details and Dockerfile.
Image | Description |
---|---|
rocker/r-base | base package to build from |
rocker/r-devel | base plus R-devel from SVN |
rocker/rstudio | base plus RStudio Server |
rocker/hadleyverse | rstudio + Hadley’s packages, LaTeX |
rocker/ropensci | hadleyverse + rOpenSci packages |
rocker/r-devel-san | base, SVN’s R-devel and SAN |
Bioconductor have a nice page about the official docker images, please read for more details.
Image |
---|
bioconductor/devel_base |
bioconductor/devel_core |
bioconductor/devel_flow |
bioconductor/devel_microarray |
bioconductor/devel_proteomics |
bioconductor/devel_sequencing |
bioconductor/release_base |
bioconductor/release_core |
bioconductor/release_flow |
bioconductor/release_microarray |
bioconductor/release_proteomics |
bioconductor/release_sequencing |
To understand the image quickly here is the short instruction for the image name:
rocker/rstudio
rocker/rstudio-daily
Dockerhub also provide public/private repos, you can search existing tools without building yourself, it’s very likely some popular tool already have docker container well maintained there.
Tutorial comming soon.
Example Seven Bridges registry:
images.sbgenomics.com/<repository>[:<tag>]
cgc-images.sbgenomics.com/<repository>[:<tag>]
Our goal here is to making a cwl app to generate unif random numbers, yes, the core function is runif()
, it’s native function in R.
runif
## function (n, min = 0, max = 1)
## .Call(C_runif, n, min, max)
## <bytecode: 0x676ad60>
## <environment: namespace:stats>
runif(10)
## [1] 0.8176919 0.7184033 0.8866669 0.4197080 0.4088952 0.7401310 0.8675907
## [8] 0.3240680 0.5995706 0.7655255
In R, we also have a nice implementation in a package called docopt
, developed by Edwin de Jonge. Check out its tutorial on github.
So let’s quickly create a command line interface for our R scripts with a dummy example. Let’s turn the uniform distribution function runif
into a command line tool.
when you check out the help page for runif
, here is the key information you want to mark down.
Usage
runif(n, min = 0, max = 1)
Arguments
n
number of observations. If length(n) > 1, the length is taken to be the number required.
min, max
lower and upper limits of the distribution. Must be finite.
I will add one more parameter to set seed, here is the R script file called runif.R
.
At the beginning of the commadn line script, I use docopt standard to write my tool help.
'usage: runif.R [--n=<int> --min=<float> --max=<float> --seed=<float>]
options:
--n=<int> number of observations. If length(n) > 1, the length is taken to be the number required [default: 1].
--min=<float> lower limits of the distribution. Must be finite [default: 0].
--max=<float> upper limits of the distribution. Must be finite [default: 1].
--seed=<float> seed for set.seed() function [default: 1]' -> doc
library(docopt)
Let’s first do some testing in your R session before you make it a full functional command line tool.
docopt(doc) #with no argumetns provided
## List of 8
## $ --n : chr "1"
## $ --min : chr "0"
## $ --max : chr "1"
## $ --seed: chr "1"
## $ n : chr "1"
## $ min : chr "0"
## $ max : chr "1"
## $ seed : chr "1"
## NULL
docopt(doc, "--n 10 --min=3 --max=5")
## List of 8
## $ --n : chr "10"
## $ --min : chr "3"
## $ --max : chr "5"
## $ --seed: chr "1"
## $ n : chr "10"
## $ min : chr "3"
## $ max : chr "5"
## $ seed : chr "1"
## NULL
Looks like it works, now let’s add main function script for this command line tool.
opts <- docopt(doc)
set.seed(opts$seed)
runif(n = as.integer(opts$n),
min = as.numeric(opts$min),
max = as.numeric(opts$max))
## [1] 0.2655087
Add Shebang at the top of the file, this is a complete example for runif.R
command line will be like this
#!/usr/bin/Rscript
'usage: runif.R [--n=<int> --min=<float> --max=<float> --seed=<float>]
options:
--n=<int> number of observations. If length(n) > 1, the length is taken to be the number required [default: 1].
--min=<float> lower limits of the distribution. Must be finite [default: 0].
--max=<float> upper limits of the distribution. Must be finite [default: 1].
--seed=<float> seed for set.seed() function [default: 1]' -> doc
library(docopt)
opts <- docopt(doc)
set.seed(opts$seed)
runif(n = as.integer(opts$n),
min = as.numeric(opts$min),
max = as.numeric(opts$max))
Let’s test this command line.
$ runif.R --help
Loading required package: methods
usage: runif.R [--n=<int> --min=<float> --max=<float> --seed=<float>]
options:
--n=<int> number of observations. If length(n) > 1, the length is taken to be the number required [default: 1].
--min=<float> lower limits of the distribution. Must be finite [default: 0].
--max=<float> upper limits of the distribution. Must be finite [default: 1].
--seed=<float> seed for set.seed() function [default: 1]
$ runif.R
Loading required package: methods
[1] 0.2655087
$ runif.R
Loading required package: methods
[1] 0.2655087
$ runif.R --seed=123 --n 10 --min=1 --max=100
Loading required package: methods
[1] 29.470174 79.042208 41.488715 88.418723 94.106261 5.510093 53.282443
[8] 89.349485 55.592066 46.204859
For full example you can check my github example
For advanced users, please read another tutorial “Creating Your Docker Container and Command Line Interface (with docopt)”, “docopt” is more formal way to construct your command line interface, but there is a quick way to make command line interface here using just commandArgs
Suppose I already have a R script like this using position mapping the arguments
fl <- system.file("docker/sevenbridges/src", "runif2spin.R", package = "sevenbridges")
cat(readLines(fl), sep = '\n')
#'---
#'title: "Uniform randome number generator example"
#'output:
#' html_document:
#' toc: true
#'number_sections: true
#'highlight: haddock
#'---
#'## summary report
#'
#'This is a randome number generator
#+
args <- commandArgs(TRUE)
r <- runif(n = as.integer(args[1]),
min = as.numeric(args[2]),
max = as.numeric(args[3]))
head(r)
summary(r)
hist(r)
Ignore the comment part, I will introduce spin/stich later. My base command will be somethine like
Rscript runif2spin.R 10 30 50
I just describe my tool in this way
library(sevenbridges)
library(readr)
fd <- fileDef(name = "runif.R",
content = read_file(fl))
rbx <- Tool(id = "runif",
label = "runif",
hints = requirements(docker(pull = "rocker/r-base"),
cpu(1), mem(2000)),
requirements = requirements(fd),
baseCommand = "Rscript runif.R",
stdout = "output.txt",
inputs = list(input(id = "number",
type = "integer",
position = 1),
input(id = "min",
type = "float",
position = 2),
input(id = "max",
type = "float",
position = 3)),
outputs = output(id = "random", glob = "output.txt"))
Now copy-paste the json into your project app and run it in the cloud to test it
How about named argumentments? I will still recommend use “docopt” package, but for simple way.
fl <- system.file("docker/sevenbridges/src", "runif_args.R", package = "sevenbridges")
cat(readLines(fl), sep = '\n')
Warning in readLines(fl): incomplete final line found on '/tmp/RtmptCGgRL/
Rinst3bc539754275/sevenbridges/docker/sevenbridges/src/runif_args.R'
#'---
#'title: "Uniform randome number generator example"
#'output:
#' html_document:
#' toc: true
#'number_sections: true
#'highlight: haddock
#'---
#'## summary report
#'
#'This is a randome number generator
#+
args <- commandArgs(TRUE)
## quick hack to split named arguments
splitArgs <- function(x){
res <- do.call(rbind, lapply(x, function(i){
res <- strsplit(i, "=")[[1]]
nm <- gsub("-+", "",res[1])
c(nm, res[2])
}))
.r <- res[,2]
names(.r) <- res[,1]
.r
}
args <- splitArgs(args)
#+
r <- runif(n = as.integer(args["n"]),
min = as.numeric(args["min"]),
max = as.numeric(args["max"]))
summary(r)
hist(r)
write.csv(r, file = "out.csv")
Rscript runif_args.R --n=10 --min=30 --max=50
I just describe my tool in this way, note, I use separate=FALSE
and add =
to my prefix as a hack.
library(readr)
fd <- fileDef(name = "runif.R",
content = read_file(fl))
rbx <- Tool(id = "runif",
label = "runif",
hints = requirements(docker(pull = "rocker/r-base"),
cpu(1), mem(2000)),
requirements = requirements(fd),
baseCommand = "Rscript runif.R",
stdout = "output.txt",
inputs = list(input(id = "number",
type = "integer",
separate = FALSE,
prefix = "--n="),
input(id = "min",
type = "float",
separate = FALSE,
prefix = "--min="),
input(id = "max",
type = "float",
separate = FALSE,
prefix = "--max=")),
outputs = output(id = "random", glob = "output.txt"))
Alternative, you can use spin/stich from knitr to generate report directly from a Rscript with special format. For example, let’s use above example
fl <- system.file("docker/sevenbridges/src", "runif_args.R", package = "sevenbridges")
cat(readLines(fl), sep = '\n')
Warning in readLines(fl): incomplete final line found on '/tmp/RtmptCGgRL/
Rinst3bc539754275/sevenbridges/docker/sevenbridges/src/runif_args.R'
#'---
#'title: "Uniform randome number generator example"
#'output:
#' html_document:
#' toc: true
#'number_sections: true
#'highlight: haddock
#'---
#'## summary report
#'
#'This is a randome number generator
#+
args <- commandArgs(TRUE)
## quick hack to split named arguments
splitArgs <- function(x){
res <- do.call(rbind, lapply(x, function(i){
res <- strsplit(i, "=")[[1]]
nm <- gsub("-+", "",res[1])
c(nm, res[2])
}))
.r <- res[,2]
names(.r) <- res[,1]
.r
}
args <- splitArgs(args)
#+
r <- runif(n = as.integer(args["n"]),
min = as.numeric(args["min"]),
max = as.numeric(args["max"]))
summary(r)
hist(r)
write.csv(r, file = "out.csv")
You command is something like this
Rscript -e "rmarkdown::render(knitr::spin('runif_args.R', FALSE))" --args --n=100 --min=30 --max=50
And so I describe my tool like this with docker image rocker/hadleyverse
this contians knitr and rmarkdown package.
library(readr)
fd <- fileDef(name = "runif.R",
content = read_file(fl))
rbx <- Tool(id = "runif",
label = "runif",
hints = requirements(docker(pull = "rocker/hadleyverse"),
cpu(1), mem(2000)),
requirements = requirements(fd),
baseCommand = "Rscript -e \"rmarkdown::render(knitr::spin('runif.R', FALSE))\" --args",
stdout = "output.txt",
inputs = list(input(id = "number",
type = "integer",
separate = FALSE,
prefix = "--n="),
input(id = "min",
type = "float",
separate = FALSE,
prefix = "--min="),
input(id = "max",
type = "float",
separate = FALSE,
prefix = "--max=")),
outputs = list(output(id = "stdout", type = "file", glob = "output.txt"),
output(id = "random", type = "file", glob = "*.csv"),
output(id = "report", type = "file", glob = "*.html")))
You will get a report in the end
We cannot really make a Rmarkdown file executable in it by simply put
#!/bin/bash/Rscript
In your markdown
Of course, we can figure out a way to do it in liftr
or knitr
. But rmarkdown allow you to pass parameters to your Rmardown template, please read this tutorial Parameterized Reports. This doesn’t solve my problem that I want to directly describe command line interface in the markdown template. However, here is alternative method:
Create an command line interface to pass params
from docopt into rmarkdown::render()
function. In this way, we can pass as many as possible parameters from command line interface into our Rmarkdown template.
So here we go, here is updated methods and it’s also what I use for another tutorial about RNA-seq workflow.
fl <- system.file("docker/sevenbridges/src/", "runif.R", package = "sevenbridges")
Here is the current content of command line interface
#!/usr/local/bin/Rscript
'usage: runif.R [--n=<int> --min=<float> --max=<float> --seed=<float>]
options:
--n=<int> number of observations. If length(n) > 1, the length is taken to be the number required [default: 1].
--min=<float> lower limits of the distribution. Must be finite [default: 0].
--max=<float> upper limits of the distribution. Must be finite [default: 1].
--seed=<float> seed for set.seed() function [default: 1]' -> doc
library(docopt)
opts <- docopt(doc)
## create param list
lst <- list(n = as.integer(opts$n),
min = as.numeric(opts$min),
max = as.numeric(opts$max),
seed = as.numeric(opts$seed))
## execute your Rmarkdown with these parameters
rmarkdown::render("/report/report.Rmd", BiocStyle::html_document(toc = TRUE),
output_dir = ".", params = lst)
And here is the report template
---
title: "Uniform randome number generator example"
output:
BiocStyle::html_document:
toc: true
number_sections: true
highlight: haddock
css: style.css
includes:
in_header: logo.md
params:
seed: 1
n: 1
min: 0
max: 1
---
```{r style, echo = FALSE, results = 'asis'}
BiocStyle::markdown(css.files = "custom.css")
```
## Summary
```{r}
set.seed(params$seed)
r <- runif(n = as.integer(params$n),
min = as.numeric(params$min),
max = as.numeric(params$max))
summary(r)
hist(r)
```
To make things more reproducible and explicit and automatic, you can do a autohook to automatically build your container/image on docker hub. Here is what I do
There are more examples under ‘inst/docker’ folder, you can check out how to describe command line and build docker, how to make report template. You can read the online github code. Or you can read another tutorial about how we wrap RNA-seq workflow from bioconductor.