write_prov: Write a provenance trace into JSON-LD

In cboettig/prov: A Semantic Data Provenance Index

Write a provenance trace into JSON-LD

Description

Write a provenance trace into JSON-LD

Usage

write_prov(
  data_in = NULL,
  code = NULL,
  data_out = NULL,
  meta = NULL,
  creator = NULL,
  title = NULL,
  description = NULL,
  issued = as.character(Sys.Date()),
  license = "https://creativecommons.org/publicdomain/zero/1.0/legalcode",
  provdb = "prov.json",
  append = TRUE,
  schema = c("http://schema.org", "http://www.w3.org/ns/dcat"),
  embed_actions = is.null(code),
  ...
)

Arguments

data_in	path or URI for input data
code	path or URI for code
data_out	path or URI to output data
meta	path or URI to metadata describing the workflow
creator	URI, list node, or text for creator
title	Dataset title, character string
description	Dataset description, character string
issued	publication date, as Date or character object
license	URL to a copyright license
provdb	path to output JSON file, default "prov.json"
append	Should we append to existing json or overwrite it?
schema	Use schema.org or DCAT2 schema? See details.
embed_actions	should we incldue schema:Action to create?
...	additional named elements passed to Dataset

Details

If creator, title, and description are all empty, will serialize only a graph of distribution (data download) elements, not a Dataset.

Additional elements passed through ... must be explicitly namespaced, e.g. dcat:version, when using DCAT2 schema. When using schema.org, elements must be in schema.org namespace.

Provenance can be expressed in (purely) schema.org or as DCAT2 (includes terms from DCTERMS, PROV, DCAT2, CITO ontologies). The latter is more expressive in terms of provenance. Also note DCAT2 but not schema.org can explicitly encode compression and metadata file relationships.

Examples

 
## Use temp files for illustration only
provdb <- tempfile(fileext = ".json")
input_data <- tempfile(fileext = ".csv")
output_data <- tempfile(fileext = ".csv")
code <- tempfile(fileext = ".R")

## A minimal workflow: 
write.csv(mtcars, input_data)
out <- lm(mpg ~ disp, data = mtcars)
write.csv(out$coefficients, output_data)

# really this would already exist...
writeLines("out <- lm(mpg ~ disp, data = mtcars)", code)

## And here we go: 
write_prov(input_data, code, output_data, provdb = provdb,  
           append= FALSE)
 
## Include a title to group these into a Dataset:
write_prov(input_data, code, output_data, provdb = provdb,
           title = "example dataset with provenance",  append= FALSE)
           

cboettig/prov documentation built on Feb. 12, 2023, 5:54 p.m.