The R script runs without problems, but the target pipeline gives error

Question

I have started using targets (from package(targets)) in order to avoid re-runing my analyses several times and make them more reproducible. However, the same analysis that works fine when R in a normal R script gives me errors when I run the run.R file (using targets::tar_make()). I am using the data and script from this course: https://biostats-r.github.io/biostats/targets/

The data can be obtained using:

usethis::use_course("biostats-r/targets_workflow_svalbard")

My _targets.R file looks like this:

# Load packages required to define the pipeline:
library(targets)

# Set target options:
tar_option_set(
  packages = c("tibble", "tidyverse")
)

# Run the R scripts in the R/ folder with your custom functions:
tar_source("R/functions2.R")

# Replace the target list below with your own:
list(
  tar_target(
    name = raw_traits,
    command = "data/PFTC4_Svalbard_2018_Gradient_Traits.csv",
    format = "file"
  ),
  tar_target(
    name = traits,
    command = clean_data(raw_traits)
  ),
  tar_target(
    name = mod_area,
    command = fit_model(data = traits,
                        response = "Value",
                        predictor = "Gradient")
  ),
  tar_target(
    name = fig_area,
    command = make_figure(traits)
  )
)

And myy functions2.R file is as follows:

# clean data
clean_data <- function(raw_traits){
  traits <- raw_traits |>
    filter(!is.na(Value)) |>
    # order factor and rename variable gradient
    mutate(Gradient = case_match(Gradient,
                                 "C" ~ "Control",
                                 "B" ~ "Nutrients"),
           Gradient = factor(Gradient, levels = c("Control", "Nutrients"))) |>
    filter(Taxon == "alopecurus magellanicus",
           Trait == "Leaf_Area_cm2")
}

# run a linear regression
fit_model <- function(data, response, predictor){
  mod <- lm(as.formula(paste(response, "~", predictor)), data = data)
  mod
}

# make figure
make_figure <- function(traits){
  ggplot(traits, aes(x = Gradient, y = Value)) +
    geom_boxplot(fill = c("grey80", "darkgreen")) +
    labs(x = "", y = expression(Leaf~area~cm^2)) +
    theme_bw()
}

When I run targets:tar_make() I get the following error:

✔ skip target raw_traits
▶ start target traits
✖ error target traits
▶ end pipeline [0.61 seconds]
Error:
! Error running targets::tar_make()
  Error messages: targets::tar_meta(fields = error, complete_only = TRUE)
  Debugging guide: https://books.ropensci.org/targets/debugging.html
  How to ask for help: https://books.ropensci.org/targets/help.html
  Last error: no applicable method for 'filter' applied to an object of class "character"

I have tried using only one function to make it simpler, but I always end up having a similar problem. Any suggestions on what the problem is, and how I could solve it? I find it strange that the script runs fine by itself, but then gives errors when using targets.

Answer 1

I got help from a colleague and found out the answer. I write it here in case someone ends up with the same problem. I just needed to load the file as a different target. It is not enough to define the file, you also need to load it. The code would then look like:

list(
  tar_target(
    name = file,
    command = "data/PFTC4_Svalbard_2018_Gradient_Traits.csv",
    format = "file"
  ),
  tar_target(
    name = raw_traits,
    command = read_csv(file)
  tar_target(
    name = traits,
    command = clean_data(raw_traits)
  ),
  ...