R visualizations defined by mine admins

This issue outlines an idea that may be worth pursuing in the future. It's an evolution of the *R playground* idea, which instead focused on end-users writing R code to visualize the results of an im-table.

The greatest drawback of the current Tool API lies in its flexibility; the fact that each tool is a complete JS application, requiring its authors to be familiar with the modern JS stack, and learn new plotting libraries. They also tend to be mine-specific (although there are ways to improve this), which puts the burden on each mine admin to develop their own Tools.

There are lots of potential [bridges](https://github.com/scicloj/clojisr) between Bluegenes' backend and R. If combined with the new `bluegenes-properties` webservice, we could allow mine admins to define visualizations using R code with a pre-initialised dataframe based on the report page or lists analysis data, and ggplot2 (possibly other libraries). This R code would then be persisted and run for all users alongside existing visualizations (as only mine admins can define the R code, we *might* be able to sidestep the problem of sandboxing for security).

The tradeoff of the R visualizations is that they'll likely be static SVG or images. Introducing interactivity may be possible with R shiny or [plotly R](https://plotly-r.com/index.html) (places more load on the R backend) or [other tools](https://davidgohel.github.io/ggiraph/articles/offcran/using_ggiraph.html) for adding JS interactivity (usually in the form of tooltips) to R plots.

*Implementation notes:*
- Along with the R code as a string, we'd need to also persist an object similar to the Tool API's *config.json* which will be used to determine when the R visualization is displayed
- The R visualization will probably want to build its own query based on the database ID(s) passed from Bluegenes; for this using the existing InterMineR package might be suitable