[Developer Issue]: wrapper function to run a FIMS model

[kellijohnson-NOAA]

Issue details

It would be nice to have a single function to run a FIMS model where everything is ran in an environment inside the wrapper function so your working environment is not polluted and you do not have to make multiple calls to non-FIMS functions, i.e., MakeADFun() and optim(). This wrapper function will help make FIMS more accessible to non-power users. Good examples exist in {sdmTMB} and {glmmTMB}. This topic is touched on in the following issues

• [Feature]: calculate model start/end time, model call, and version of FIMS #617
• [Bug]: Difference in point estimates from tmb sdreport vs. tmb output object #616
• [Feature]: Run multiple instances of FIMS in the same R session #378
• [Feature]: R crashes rather than giving error when there are no estimated parameters #552
• [Feature]: Make output$obj portable #555

The timeline for the development of this R function is still underway but research is currently being conducted.

• Research what other packages do
• Implement a skeleton function
• Finish the function
• Present the function in a seaside chat during any point of development for feedback
• Write tests
• Document the function
• Update the vignette and case-studies (but maybe leave an example of how a power user would do it in so there are examples of all avenues), feel free to open an issue in the case studies repo and assign someone to modify the case study code to use the new function
• close associated issues (see above) that pertain to this function if they are finished

Notes for the development of this function can be found here on Google Drive.

[Andrea-Havron-NOAA]

The Kristensen et al. 2016 paper performed a comparison between optim and nlminb and found nlminb to be more stable and faster.

[Bai-Li-NOAA]

Thanks @Andrea-Havron-NOAA, we will use nlminb in the wrapper function as suggested.