Helper for downstream batched targets

[wlandau]

Prework

• I understand and agree to the code of conduct and contributing guidelines.
• If there is already a relevant issue, whether open or closed, comment on the existing thread instead of posting a new issue.
• New features take time and effort to create, and they take even more effort to maintain. So if the purpose of the feature is to resolve a struggle you are encountering personally, please consider first posting a "trouble" or "other" issue so we can discuss your use case and search for existing solutions first.

Proposal

tar_rep() establishes a batching scheme, but then it is up to the user to write custom code to iterate over reps within batches. If possible, it would be nice to have a helper automatically process the iterations of an individual batch. This sort of thing is so general that I am not sure it is possible Sketch:

# _targets.R
library(targets)
library(tarchetypes)
list(
  tar_rep(data1, simulate_data(), batches = 40, reps = 25),
  tar_rep(data2, simulate_data(), batches = 40, reps = 25),
  tar_rep(data3, simulate_data()),
  tar_target(analysis, tar_map_reps(analyze_data(data1, data2, data3)), pattern = map(data1, data2))
)

tar_map_reps() should automatically detect which targets are batched (e.g. data1 and data2 but not data3) and infer how to iterate over the batches given the data types (lists or data frames).

[wlandau]

I no longer like the above helper tar_map_reps(). I propose a target factory called tar_rep_map() that abstracts away all the dynamic branching and accepts multiple batched targets.

# _targets.R
library(targets)
library(tarchetypes)
list(
  tar_rep(data1, simulate_data(), batches = 40, reps = 25),
  tar_rep(data2, simulate_data(), batches = 40, reps = 25),
  tar_rep(data3, simulate_data()),
  tar_rep_map(analysis, analyze_data(data1, data2, data3), data1, data2) # Use ... to declare batched targets data1 and data2.
)

[wlandau]

Another idea: tar_batch(), a target factory to create a batching structure on an existing return value without having to go through tar_rep().

[wlandau]

Another idea: tar_batch(), a target factory to create a batching structure on an existing return value without having to go through tar_rep().

Thought about it, but tar_batch() and a more generalized tar_rep_map() is extremely brittle. Let's stick with operations directly downstream of tar_rep() targets.