Adapt to new posterior database structure

[eerolinna]

This pull request changes to new posterior database structure and adapts rpackage to the new structure.

Fixes #8

[eerolinna]

The problem with this solution is that we no longer have access to the data names and model names from po. One way to fix this would be to add

po$data_name <- po$data
po$model_name <- po$data

to posterior in posterior.R

This is maybe not the best solution still though

[eerolinna]

Another way would be not to overwrite po$model and po$data but instead make new fields like po$model_info and po$data_info

A third way would be to not attach the info file contents to po but load them on-demand. Then for example in model_code_file_path

https://github.com/MansMeg/posterior_database/blob/96dcb71330f376edecf9b7f8613c067d0d58cb85/rpackage/R/model_code.R#L12

would become something like

model_info <- load_model_info(x$model)
mcfp <- file.path(x$pdb$path,  model_info[[framework]]) 

[eerolinna]

Now this PR doesn't overwrite anything.

The slots in the posterior object are now

• model_name: from posterior json file
• data_name: from posterior json file
• model_info: contents of model info file
• data_info: contents of data info file

[eerolinna]

This should include everything that is needed now. I think you haven't added the contributor etc slots yet so I couldn't include them here. Would be great if you can first merge this and then add the new slots

[eerolinna]

I found a problem in master branch that was causing tests to fail, some dataset zip files had json files inside them with wrong filename (lowercase n instead of capital N) that caused stan_data(po) to not find the data file, I corrected that here also.

[MansMeg]

Great! I had problem finding that problem yesterday.