clean out the "models" directory in this repository

[iantaylor-NOAA]

During early stages of the lingcod work, model input and output files were committed to this repository. However, it became clear that they were adding clutter and focus shifted to sharing models only on Google Drive.

As proposed by @kellijohnson-NOAA earlier today, once we have completed archiving the assessment (#187), we should remove all but the most important models, and also delete the file history for those files. This should provide a cleaner starting point for future assessors.

Question: does it make sense to keep ANY model files in the repository? They will all be on Google Drive, on our local computers, and backed up on NWFSC servers. The base model and states of nature will also be in the NWFSC archive.

[brianlangseth-NOAA]

It would be cleanest to have one location. Since it appears we've already chosen the Drive location to be that location I think removing all model files from the repository is the cleanest. Presumably, for the package to work, one would need to download all of the model files from the Drive to the "models" folder? If so, we could add a txt file saying as such, which could help out our future selves, as well as anyone else trying to work with the package.

[iantaylor-NOAA]

That makes sense. We already have a models/README.md file so I can just add the following text:

The model files are not stored in the repo and instead located on users
local computers and shared via Google Drive:
https://drive.google.com/drive/folders/1xkIZznTkVXHjhjWJOYJrMQO93qfD-O3V
Please contact @iantaylor-NOAA or @kellijohnson-NOAA for access if you need it.

Some complicating factors that I hadn't thought about before are

1. the /models/ folder is used to share figures and tables that get used in the assessment, such as the profile plots for the final base models: https://github.com/iantaylor-NOAA/Lingcod_2021/tree/main/models/2021.s.018.001_fixTri3_profile_SR_LN(R0)_prior_like_1
   to build the doc, you need these plots, but you don't need to have the SS3 input or output files (they are only on @kellijohnson-NOAA's computer because she was the one who ran nwfscDiags). In that case it probably doesn't make sense to upload the extra model files to Google Drive, but does it make sense to delete the figs from the repo?
2. there's no equivalent of .gitignore for Google Drive, so it will take some work to only upload useful files. This would include
   • running a batch script to clean out the junk ADMB files that we'll never be interested in
   • being selective in what we upload to Google Drive (e.g. the subdirectories tables, plots, and custom_plots for the final base models, but not plots_Feb2_CAPAM_talk or plots_old or tex_tables)

[kellijohnson-NOAA]

I think there is a difference in what we want for the current version to work and what we would want going forward. I want to focus what remains in this repository on the latter rather than former and I believe we need the "base model" input files for each area.

[kellijohnson-NOAA]

@iantaylor-NOAA the sensitivities can be easily recreated; so, I think that we should just delete the figures from the repo. Regarding the .gitignore for Google Drive, are you thinking about for models now or for going forward in the future?