Final tips

[github-learning-lab[bot]]

You are nearly done with the second pipelines course! We have a few final thoughts before you level up your pipelines capability :trophy:

---

First, you may have seen a folder called _targets :file_folder: created after running tar_make() for the first time. This folder contains the intermediate objects built when running the pipeline, as well as, status metadata about the most recent build for each target. This kind of information is extremely important to you on your computer, but your collaborators will have unique status information on their own computer. For this reason, make sure to add _targets/* to your .gitignore file so that it won't be committed.

We encourage excessive commenting of code! It is so important to explain what you are doing and why you are doing it, especially when you come up with a really clever line of code :bulb:. With targets, any commented lines of code (ones beginning with #) you add to your functions will not be seen as a change to the file and will not trigger a rebuild of targets. So comment away freely, regardless of whether it's before or after you've written and run a function!

Final tip, although a targets pipeline is written and executed in R, you can still call functions from another computing language or even invoke system commands :sunglasses:. For example, maybe you need to read in a NumPy file that was written from a Python workflow :snake:. You can use reticulate and targets to read in this file like so:

np <- reticulate::import('numpy')

p1_targets_list = list(
  ...
  tar_target(p1_obs_data_npy, '1_data/in/obs_data.npy', format='file'),
  tar_target(
    p1_obs_data,
    np$load(p1_obs_data_npy)
  )
)

Or maybe you wrote your own custom Python function and you want to use it in your targets pipeline. Great! Just import your function(s) and use them in your pipeline:

# import custom Python functions 
prep_data_py = reticulate::import_from_path(module = 'prep_data', path = '2_model_prep/src/')

p2_targets_list = list(
  ...
  tar_target(
    p2_data_munged_npz,
    prep_data_py$prep_data(
      obs_file = p1_obs_data_npy,
      locations = p2_config[['model_loc']],
      start_model = p2_config[['start']],
      stop_model = p2_config[['stop']]),
    format = 'file'
  )
)

---

targets is part of the rOpenSci community of packages and is being used by a wide net of R users. If you have further questions, you can always consult a Data Science colleague, but also try Google and take advantage of the extensive documentation available in the targets R Package User Manual.

Close this issue to continue

[github-learning-lab[bot]]

---

When you are done poking around, check out the next issue.