Add D-M, random seed argument, and code improvements

[Cole-Monnahan-NOAA]

Edited to add point 5.

This PR does five things:

1. Adds functionality for the Dirichlet-multinomial distribution via the new argument theta.
2. Adds the random seed as an argument so that users can try different ones to see the impact of stochasticity.
3. Minor improvements to workflow, namely forcing the order of the input list and columns to match (i.e., no reordering alphabetically), and updating saving files in plot_osa so that the user has more control with the plot argument.
4. Fixed package documentation (NAMESPACE).
5. Added 95% confidence interval (via addCI argument) and more flexible placement of the label on QQ plots (via hjust and vjust)

[JaneSullivan-NOAA]

@Cole-Monnahan-NOAA do you have a D-M example we could add? maybe to the existing GOA pk script?

[Cole-Monnahan-NOAA]

Hi Jane, I didn't make one. Perhaps we should consider moving away from the packaged examples and just use simulated data for both the MN and DM. That'd be easier on the user side. We could keep the existing scripts and perhaps they'd be better used as tests for when we make changes. If you approve this idea I'll update the PR.

[JaneSullivan-NOAA]

i agree we need some tests. i can see the benefit of using simulated data sets. i had used real life examples so that new users could follow a workflow that most closely resembled their own model (SS3, AMAK, or some other bespoke ADMB model). but, your point is valid i think.

[Cole-Monnahan-NOAA]

@JaneSullivan-NOAA I signed myself up for that idea in this issue but suggest we have that as a separate PR to keep things a little more organized. So I'll do another PR once this one is done and merged in. Any objections?

[Cole-Monnahan-NOAA]

Done with this commit:
3373eda

Also moved the position and allow the user to adjust that via arguments vjust and hjust.