Add integration tests for convert_output()

[Bai-Li-NOAA]

Ensure that the function works with a set of input files and returns standardized assessment model results

[Bai-Li-NOAA]

@Schiano-NOAA, I've added integration tests for convert_output() using the SS3 test models (see test-convert-output branch). These tests are set to run on three operating systems. I have a couple of questions before merging the changes into main:

• When testing with SS3 model output files, five of them triggered the following warning messages:

Warning messages:
1: In convert_output(output_file = "Report.sso", outdir = all_models[i],  ... :
  There are multiple reported error metrics.
2: In convert_output(output_file = "Report.sso", outdir = all_models[i],  ... :
  There are multiple reported error metrics.
3: In convert_output(output_file = "Report.sso", outdir = all_models[i],  ... :
  There are multiple reported error metrics.
4: In convert_output(output_file = "Report.sso", outdir = all_models[i],  ... :
  There are multiple reported error metrics.
5: In convert_output(output_file = "Report.sso", outdir = all_models[i],  ... :
  There are multiple reported error metrics.

Did I possibly misconfigure any arguments? Or is this behavior expected, and should we be concerned about those error metrics? If not, I'll modify the tests to expect warnings when running convert_output().

• I noticed that when file_save = TRUE, the formatted object is saved as a CSV in the specified directory but not saved in the environment. Would it be possible to both save the object in the environment and write it to a local file? This way, users could pass the object to other report generation functions while still saving a local copy. Or, was there a specific design reason behind this?

[Schiano-NOAA]

@Bai-Li-NOAA I made the warning to let the user know about some minor changes in the conversion process. These are nothing to worry about and I will most likely make changes to this in the future.

For the file_save = TRUE. My intention was that if file_save = TRUE then the user will be exporting it as a csv otherwise when file_save = FALSE, the user can name the function then save it as an object to their global environment such as

output <- convert_output(file_save = FALSE)

Setting it up this way allows the user either option. I could just set it up so it does both, but I feel like giving the user the option to save as an object allows them to manipulate it and then the user could just save it as a csv themself after the fact if they want to.

[Bai-Li-NOAA]

@Schiano-NOAA, great! I've modified the tests so that warnings no longer cause a test failure. As for the file_save flag, the original setting works perfectly. My bad, I made a mistake on my end and ended up with NULL when running output <- convert_output(file_save = FALSE). Will submit a pull request once all the tests pass.