In bvancil/tricordr: Integrates Process Provenance and Instrumentation to Pipelines of Tidyverse Primatives

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

tricordr

The goal of tricordr is to integrate process provenance and instrumentation to pipelines of tidyverse primatives

Installation

You can install the released version of tricordr from CRAN with:

install.packages("tricordr")

And the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("bvancil/tricordr")

Example

Warning: None of this actually works yet.

Eventually, we want full provenance, but this is sort of what we're going for.

We'll start by creating some test data.

library('dplyr')
library('tibble')
library('tricordr')

# Here's some test data
data_size <- 100L
test_data <- tibble::tibble(
  x = base::sample(c(0L, 1L), data_size, replace = TRUE),
  y = base::sample(base::seq(0L, 2L), data_size, replace = TRUE)
)

In our data pipeline, we want to track what happens to our data. For instance, we might want to add another variable and transform the others.

final_data <- test_data %>% 
  dplyr::mutate(z = x * y, y1 = y, y = x, x = 2L - y1)

What happened? It will be tricky to figure out later. Instead, we can use tricordr to decorate the dplyr::mutate function so that we keep track.

test_provenance <- tricordr::Provenance$new()
test <- test_provenance$wrap_operations()

# Now we change `dplyr::mutate` to `test$mutate`
final_data <- test_data %>% 
  test$mutate(z = x * y, y1 = y, y = x, x = 2L - y1)

print(test_provenance)

We can create difference provenances for different pipelines and combine them later.

Code of conduct

Please note that the 'tricordr' project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.

Similar work

• {tidylog}

bvancil/tricordr documentation built on Jan. 24, 2020, 1:25 a.m.