What improvements would you make to the process as a whole or individual parts of the workflow?

[Schiano-NOAA]

Indicate your answer to this question after completing the activity.

[BenWilliams-NOAA]

When using the create_tables_doc and create_figures_doc functions the results have a hardcoded rda_dir at the top of each .qmd - this is a non-starter for working with collaborators. I think that everything in this package will work better if the user is forced into a project framework where the path root is implicitly the top-level of the project. This is most easily achieved using rprojroot or the here package(s) - not setwd() and getwd() - though the latter isn't so bad.

Additionally, rda_dir has to go - should be replaced with figs and tables. These folder names are more clear and assessments often have many tables and many figures - putting them all in one place is a mess and can make it hard to find specific files.

[BenWilliams-NOAA]

In the "#5. Add standard output to template" - the output is quite large and takes a while to save using write.csv, I suggest recommending the use of vroom::vroom_write or data.table::fwrite to substantially speed things up.

On my machine
write.csv takes ~ 33.8 seconds
vroom takes ~ 273 milliseconds
fwrite ~ 91 milliseconds

similar, though less extreme results when reading in a .csv

[kellijohnson-NOAA]

nit: the messages upon loading each of the libraries are not needed, I think almost everyone should know where to go to submit issues and as you work forward through the package you are not going to want people to submit issues for things regarding not knowing how to use the package, those will be user questions. I think it is better to just reiterate to people verbally and in documentation that issues can be filed for bugs and features and that discussions can be used for learning and ideas. Essentially, this message upon loading is not standard and when I saw it I thought it would provide non-trivial information and instead it just clutters up the console.

[kellijohnson-NOAA]

convert_output() has 7 input arguments

convert_output <- function(
    output_file = NULL,
    outdir = getwd(),
    model = NULL,
    fleet_names = NULL,
    file_save = FALSE,
    savedir = getwd(),
    save_name = "std_model_output") {

but really it only needs two, output_file and fleet_names. I do not understand the need for directory arguments along with file-name arguments. Why cannot you just pass the full file path to output_file? Also, I am pretty sure that you should be able to tell what model the file is from just using an internal function that runs some checks rather than requiring the user to specify. I do not think that the function should include any ability to save the file. The output object can easily be saved by the user after the fact.