Pass the list of reactive data to the module.

[gogonzo]

• Blocked by Change constructor of the FilteredData to not depend on TealData teal.slice#33

Our objective is teal_module(s) to receive the data as a list(adsl = reactive(<data.frame>), x = reactive(<data.frame>)) instead of taking data from the FilteredData object. In the flow-chart below what we need to do in this issue is distinguished by "bolded" frames. This means that if teal_module has a data argument, we need to pass a list which has expr, join_keys and metadata as attributes.

tm_module(
  a b c <instructions for choices> # currently data_extract
  data <list of reactives, optionally with expr, metadata and join_keys attr>, 
  # datasets - FilterPanel-api if the communication with filteredData needed
  # reporter - if reporter-api needed 
)

Practically my first guess is that following lines might be enough before calling the module - but it depends how far refactor of the FilteredData will go (insightsengineering/teal.slice#33)

data <- lapply(
  <module datanames>, 
  function(x) {
    dataset <- datasets$get_data(dataname = x, filtered = TRUE))
    attr(dataset, "metadata") <- <metadata>
    dataset
  }
)
attr(data, "join_keys") <- <join keys>
attr(data, "expr") <- c(datasets$get_code(<module datanames>, datasets$get_filter_expr(<datanames>))

---

Currently following methods from FilteredData are used in the modules, they should be passed somehow in the data attributes:

• $get_data(filtered = TRUE)
• $get_data(filtered = FALSE) ❌ see occurrences and here. This can be fixed by including unfiltered data to the module also.
• $datanames() - this will be the names of the list
• $get_metadata(), $get_keys() - should we pass using ...? Please find the optimal solution.
• $get_varlabels - this can be extracted from the raw data.frame object