Assertive Programming for R Analysis Pipelines.

assertr

What is it?

The assertr package supplies a suite of functions designed to verify assumptions about data early in an analysis pipeline so that data errors are spotted early and can be addressed quickly.

This package does not need to be used with the magrittr/dplyr piping mechanism but the examples in this README use them for clarity.

Installation

You can install the latest version on CRAN like this

    install.packages("assertr")

or you can install the bleeding-edge development version like this:

    install.packages("devtools")
    devtools::install_github("ropensci/assertr")

What does it look like?

This package offers five assertion functions, assert, verify, insist, assert_rows, and insist_rows, that are designed to be used shortly after data-loading in an analysis pipeline...

Let’s say, for example, that the R’s built-in car dataset, mtcars, was not built-in but rather procured from an external source that was known for making errors in data entry or coding. Pretend we wanted to find the average miles per gallon for each number of engine cylinders. We might want to first, confirm

• that it has the columns "mpg", "vs", and "am"
• that the dataset contains more than 10 observations
• that the column for 'miles per gallon' (mpg) is a positive number
• that the column for ‘miles per gallon’ (mpg) does not contain a datum that is outside 4 standard deviations from its mean, and
• that the am and vs columns (automatic/manual and v/straight engine, respectively) contain 0s and 1s only
• each row contains at most 2 NAs
• each row is unique jointly between the "mpg", "am", and "wt" columns
• each row's mahalanobis distance is within 10 median absolute deviations of all the distances (for outlier detection)

This could be written (in order) using assertr like this:

    library(dplyr)
    library(assertr)

    mtcars %>%
      verify(has_all_names("mpg", "vs", "am", "wt")) %>%
      verify(nrow(.) > 10) %>%
      verify(mpg > 0) %>%
      insist(within_n_sds(4), mpg) %>%
      assert(in_set(0,1), am, vs) %>%
      assert_rows(num_row_NAs, within_bounds(0,2), everything()) %>%
      assert_rows(col_concat, is_uniq, mpg, am, wt) %>%
      insist_rows(maha_dist, within_n_mads(10), everything()) %>%
      group_by(cyl) %>%
      summarise(avg.mpg=mean(mpg))

If any of these assertions were violated, an error would have been raised and the pipeline would have been terminated early.

Let's see what the error message look like when you chain a bunch of failing assertions together.

    > mtcars %>%
    +   chain_start %>%
    +   assert(in_set(1, 2, 3, 4), carb) %>%
    +   assert_rows(rowMeans, within_bounds(0,5), gear:carb) %>%
    +   verify(nrow(.)==10) %>%
    +   verify(mpg < 32) %>%
    +   chain_end
    There are 7 errors across 4 verbs:
    -
             verb redux_fn           predicate     column index value
    1      assert     <NA>  in_set(1, 2, 3, 4)       carb    30   6.0
    2      assert     <NA>  in_set(1, 2, 3, 4)       carb    31   8.0
    3 assert_rows rowMeans within_bounds(0, 5) ~gear:carb    30   5.5
    4 assert_rows rowMeans within_bounds(0, 5) ~gear:carb    31   6.5
    5      verify     <NA>       nrow(.) == 10       <NA>     1    NA
    6      verify     <NA>            mpg < 32       <NA>    18    NA
    7      verify     <NA>            mpg < 32       <NA>    20    NA

    Error: assertr stopped execution

What does assertr give me?

• verify - takes a data frame (its first argument is provided by the %>% operator above), and a logical (boolean) expression. Then, verify evaluates that expression using the scope of the provided data frame. If any of the logical values of the expression's result are FALSE, verify will raise an error that terminates any further processing of the pipeline.

• assert - takes a data frame, a predicate function, and an arbitrary number of columns to apply the predicate function to. The predicate function (a function that returns a logical/boolean value) is then applied to every element of the columns selected, and will raise an error if it finds any violations. Internally, the assert function uses dplyr's select function to extract the columns to test the predicate function on.

• insist - takes a data frame, a predicate-generating function, and an arbitrary number of columns. For each column, the the predicate-generating function is applied, returning a predicate. The predicate is then applied to every element of the columns selected, and will raise an error if it finds any violations. The reason for using a predicate-generating function to return a predicate to use against each value in each of the selected rows is so that, for example, bounds can be dynamically generated based on what the data look like; this the only way to, say, create bounds that check if each datum is within x z-scores, since the standard deviation isn't known a priori. Internally, the insist function uses dplyr's select function to extract the columns to test the predicate function on.

• assert_rows - takes a data frame, a row reduction function, a predicate function, and an arbitrary number of columns to apply the predicate function to. The row reduction function is applied to the data frame, and returns a value for each row. The predicate function is then applied to every element of vector returned from the row reduction function, and will raise an error if it finds any violations. This functionality is useful, for example, in conjunction with the num_row_NAs() function to ensure that there is below a certain number of missing values in each row. Internally, the assert_rows function uses dplyr'sselect function to extract the columns to test the predicate function on.

• insist_rows - takes a data frame, a row reduction function, a predicate-generating function, and an arbitrary number of columns to apply the predicate function to. The row reduction function is applied to the data frame, and returns a value for each row. The predicate-generating function is then applied to the vector returned from the row reduction function and the resultant predicate is applied to each element of that vector. It will raise an error if it finds any violations. This functionality is useful, for example, in conjunction with the maha_dist() function to ensure that there are no flagrant outliers. Internally, the assert_rows function uses dplyr'sselect function to extract the columns to test the predicate function on.

assertr also offers four (so far) predicate functions designed to be used with the assert and assert_rows functions:

• not_na - that checks if an element is not NA
• within_bounds - that returns a predicate function that checks if a numeric value falls within the bounds supplied, and
• in_set - that returns a predicate function that checks if an element is a member of the set supplied. (also allows inverse for "not in set")
• is_uniq - that checks to see if each element appears only once

and predicate generators designed to be used with the insist and insist_rows functions:

• within_n_sds - used to dynamically create bounds to check vector elements with based on standard z-scores
• within_n_mads - better method for dynamically creating bounds to check vector elements with based on 'robust' z-scores (using median absolute deviation)

and the following row reduction functions designed to be used with assert_rows and insist_rows:

• num_row_NAs - counts number of missing values in each row
• maha_dist - computes the mahalanobis distance of each row (for outlier detection). It will coerce categorical variables into numerics if it needs to.
• col_concat - concatenates all rows into strings
• duplicated_across_cols - checking if a row contains a duplicated value across columns

and, finally, some other utilities for use with verify

• has_all_names - check if the data frame or list has all supplied names
• has_only_names - check that a data frame or list have only the names requested
• has_class - checks if passed data has a particular class

More info

For more info, check out the assertr vignette

    > vignette("assertr")

Or read it here.