In FAIRDataPipeline/rSimpleModel: SimpleModel

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(rSimpleModel)
library(rDataPipeline)
library(here)
library(magick)
library(pdftools)

i_am(path = "vignettes/SEIRS.Rmd")
knitr::opts_knit$set(root.dir = here())

config.yaml file

The user should write a config.yaml file containing information pertaining to the data products used in the code run. The example config.yaml file below describes a code run with inputs:

• disease/sars_cov2/SEIRS_model/parameters/static_params

These inputs are listed in the register block, meaning that they should be downloaded to the local data store from an external source, with associated metadata stored in the local registry. These inputs are automatically converted into a read block by fair run (when data products are already present in the data registry, inputs should be listed in the read block).

A code run usually also has outputs, which are listed in the write block. In the example below, our outputs are:

• disease/sars_cov2/SEIRS_model/results/model_output
• disease/sars_cov2/SEIRS_model/results/figure

SEIRSconfig.yaml:

cat(readLines(file.path(here(), 'inst/extdata/SEIRSconfig.yaml')), sep = '\n')

The submission script should call initialise() to set up the code run, then perhaps read in some data using one of the read_*() functions (for internal file formats) or link_read() (for external file formats such as csvs). The data might now be processed in some way, or a model / analysis might bw carried out, after which the results should be saved in the local data store via one of the write_*() functions or link_write(). When the code run is complete, finalise() should be called to register the all metadata with the local registry.

fair pull

Using the CLI tool, fair pull identifies any data products listed in the register field of the config.yaml. These data products are downloaded to the local data store whilst associated metadata is registered in the local registry.

fair init --ci
fair pull inst/extdata/SEIRSconfig.yaml

The local registry should now contain three data products:

1. disease/sars_cov2/SEIRS_model/parameters/static_params,
2. disease/sars_cov2/SEIRS_model/parameters/rts, and
3. disease/sars_cov2/SEIRS_model/parameters/efoi.

fair run

Again using the CLI tool, fair run performs the code run, as written in the. submission script. In preparation for this, it translates the user-written config.yaml file for use by the Data Pipeline API. Any variables / wildcards specified by the user in the config file are cross referenced with the registry, and any data products registered by fair pull are made available to read by the current code run.

fair run inst/extdata/SEIRSconfig.yaml

Outputs

path <- get_dataproduct(data_product = "SEIRS_model/results/figure/R",
                        version = "0.0.1",
                        namespace = "testing")
image_read_pdf(path)

Provenance report

get_provenance(data_product = "SEIRS_model/results/model_output/R",
               version = "0.0.1",
               namespace = "testing")

FAIRDataPipeline/rSimpleModel documentation built on July 20, 2023, 8:15 a.m.