Running tar_* with the pipeline in a subdirectory #953

TylerGrantSmith · 2022-10-07T17:08:30Z

TylerGrantSmith
Oct 7, 2022

I am trying to figure out what I am doing wrong configuration-wise. At the top level of my R project I have a _targets.yaml file:

main:
  script: data-raw/_targets.R
  store: data-raw/_targets
  reporter_make: summary

I can only get tar_make() and tar_outdated() to run if I first change my working directory setwd("./data-raw"). How can I configure the project so this step isn't required?

wlandau · 2022-10-07T18:04:57Z

wlandau
Oct 7, 2022
Maintainer

Both paths should be relative to the directory where you run tar_make(). So if you run tar_make() from the project root, it assumes there is an existing file data-raw/_targets.R at the same project root.

To debug, you could experiment with the script and store arguments of tar_make() to see what works there. The YAML controls the default output of tar_config_get(), which controls many of the default arguments of tar_make(). Then if that works, you could test that the default outputs of tar_config_get("script") and tar_config_get("store") are what you expect and file.exists() is TRUE for the script. Then if you find you need to make changes in the yaml, then tar_make() should see them.

Other than that, hard to say without a reproducible example.

2 replies

TylerGrantSmith Oct 10, 2022
Author

So it appears to be an issue when the _targets.R file is generated from an Rmd rather than a vanilla script. Here is an example:

https://github.com/TylerGrantSmith/targets-issue

If you run tar_make using the custom config it works fine and builds or skips the single target. If you run it using the main config it will show 0 targets. But, it will work if you run withr::with_dir('data-raw', tar_make())

wlandau Oct 11, 2022
Maintainer

Thanks, I think I know what you mean now. It's a bug (#957) where targets assumes you generate _targets.R (via Target Markdown) from the same directory where you call the pipeline. The Rmd file wrote a _targets.R file from the data-raw working directory, and so _targets.R tried to find the _targets_r/ folder as if it were from the same working directory. The GitHub version now fixes this by finding _targets_r/ dynamically.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Running tar_* with the pipeline in a subdirectory #953

Uh oh!

{{title}}

Uh oh!

Replies: 1 comment 2 replies

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Select a reply

Uh oh!

Running tar_* with the pipeline in a subdirectory #953

Uh oh!

TylerGrantSmith Oct 7, 2022

Replies: 1 comment · 2 replies

Uh oh!

wlandau Oct 7, 2022 Maintainer

Uh oh!

TylerGrantSmith Oct 10, 2022 Author

Uh oh!

wlandau Oct 11, 2022 Maintainer

TylerGrantSmith
Oct 7, 2022

Replies: 1 comment 2 replies

wlandau
Oct 7, 2022
Maintainer

TylerGrantSmith Oct 10, 2022
Author

wlandau Oct 11, 2022
Maintainer