Merge in LPI data from MATSS-forecasting

[ha0ye]

MATSS-forecasting has code, contributed by @pennekampster, to import data from the Living Planet Index. I suspect this might have overlap with some other time series, which we should check, per #123

It probably makes sense to bring that data importation code over to here, perhaps also with the Ward et al. 2014 datasets.

One thing to watch out for are the minor issues with the rlpi package - we should contact them to see if that is the best source, contributing fixes, etc.

[ha0ye]

I've contacted @robinfreeman by e-mail, so hopefully we hear back soon about how to proceed on this.

[ha0ye]

No response yet, as of 2020-01-30

[ha0ye]

See Zoological-Society-of-London/rlpi#6

[ha0ye]

The preferred method for getting LPI data seems to be:
http://www.livingplanetindex.org/projects?main_page_project=LivingPlanetReport&home_flag=1

Redistribution of the dataset (as this project likely counts as) requires prior permission. I don't know how that would fit in with the procedures for adding a dataset to https://github.com/weecology/retriever . If it doesn't work to add LPI data to retriever, then the strategy seems to be to let MATSSforecasting keep its own code for getting LPI data (and update it to point to the new source), since that is a self-contained analysis project and not data redistribution.

@henrykironde, @ethanwhite What do you all think?

[ethanwhite]

There's a live link to the data on the website (http://www.livingplanetindex.org/documents/LPR2018data_public.csv) so we can add it. The whole design of the retriever is intended to avoid issues with redistribution (the user downloads the data themselves). That said, we want to be careful about notifying users of data use restrictions. At the moment I think the best way to handle this would be to put this in the license field with something like "Custom (academic only, no redistribution): http://www.livingplanetindex.org/documents/data_agreement.pdf" and make sure that this is getting reported when running retriever license.

[henrykironde]

We do have a great way to create the data package.
Download the files to new directory.
Name the directory with the name desired for the package.
Run retriever autocreate -d directory-path.
Edit the created json by populating the fields