In tengfei/rabix.R: R utilities for rabix.

BiocStyle::markdown(css.files = "custom.css")

code <- function(...) {
    cat(paste(..., sep = "\n"))
}

code2 <- function(...) {
    cat(paste("```markdown", ..., "\n", "```", sep = "\n"))
}

Introduction

Common workflow languange is a community wise effor to create specifications that enable reproducible, portable data analysis flow. This R package is based on draf2, full details are described in the website.

This package provides an R interface to create JSON/YAML file to be used in other implementation, for example, rabix provides very handy interface to desctibe your tool and flow with graphical editor to make development so much easier, it also has local executor as well.

Class

This package defines a set of S4/Reference class that represent data model and tool/flow model specified in draft2, in order to enable object-oriented programming in R. Identical class name and fields names in the CWL specification are used, and manuals are directly copied to each class to make reference checking more easier.

The top class is CWL class, it comes with soeme utility function that convert the object into list, JSON, YAML.

Constructor

To construct the object, you can simply check out the class name (same as the generator name) in the terminal or check its manual to see what fields it need.

library(rabix)
InputParameter

Then construct it with the constructor directly

ip <- InputParameter(id = "bam", type = "File", streamable = TRUE, 
               label = "Bam fles", default = "./test.bam")

Methods

show method

The first thing you need to pay attention to is that, show method is overrided, by default it output YAML which is easier to read. It automatically parse it recurdively.

ip #same as ip$show()

You can also show it as JSON

ip$show("JSON") #same as ip$show()

The default can always be called

showDefault(ip)

Convert to list

This packaeg defines a S4 method called asList, you can use it for any object, and it's different from as.list. asList automatically parse any object recursively to turn S4, R5 reference object or simple object into a list, it only parses the fields or slots in the object.

A <- setClass("A", slots = list(x = "character", y = "numeric"))
a <- A(x = "a", y = 3)
## cannot run as.list(a)
asList(a)
asList(ip) # same as ip$toList()

And CWL class extends to almost all classes defined in this pacakge, so they all comes with the method obj$toList()

Convert to YAML

To convert an object into YAML is also as easy as list, just use S4 method asYAML for any object or obj$toYAML() for CWL extended object.

ip$toYAML()

Convert to JSON

To convert an object into JSON is also as easy as list, just use S4 method asYAML for any object or obj$toJSON() for CWL extended object.

ip$toJSON()
ip$toJSON(pretty = TRUE)

A CommandLineTool example

Here I let's work through an example for a command line

samtools sort -o output.bam -l 1

And let's expose the -l -o interface into the CWL object.

Define input parameteres list

## long form
ipl <- InputParameterList(
    InputParameter(id = "bam",
                   type = "File",
                   label = "Bam file",
                   description = "Input bam file",
                   inputBinding = CommandLineBinding(
                       position = 1L,
                       separate = TRUE
                   )),
    InputParameter(id = "level",
                   type = "Integer",
                   label = "Compression Level",
                   description = "Set compression level, from 0 (uncompressed) to 9 (best)",
                   inputBinding = CommandLineBinding(
                       position = 2L
                   )),
    InputParameter(id = "prefix",
                   type = "String",
                   label = "Prefix",
                   description = "Write temporary files to PREFIX.nnnn.bam",
                   inputBinding = CommandLineBinding(
                       position = 3L
                   ))
)

Define output parameter list

opl <- OutputParameterList(
    CommandOutputParameter(
        id = "sorted",
        type = "File",
        outputBinding = CommandOutputBinding(
            glob = "*.bam"
        )
    )
)

Wrap them into CommandLineTool object

cwl <- CommandLineTool(id = "samtools-sort",
                label = "Samtools sort subcommand",
                description = "Samtools sort: sort bam into sorted bam : )",
                baseCommand = "samtools sort",
                arguments = "out.bam",
                inputs = ipl,
                outputs = opl)

Let's see it in JSON

cwl$toJSON(pretty = TRUE)

Rabix package extends this package, with it's own specific specs, and has shorthands and an easier interface to describe tools. And the web interface allow you to create/describe tools/workflows much more easier. You can also simply import an JSON file to the interface. Please check the rabix package for more details.

tengfei/rabix.R documentation built on May 31, 2019, 8:34 a.m.