Improve Documentation

[Aklakan]

There are so many features currently either badly documented or just undocumented:

• [x] graphql
• [x] geoindex option
• [x] multiple engine support
• [x] function macros
• [ ] in-memory/disk-based caching
• [ ] spark-based named graph querying (rpt sansa ngs)
• [ ] from enhancer and service enhancer - how to use them
• [ ] rmltk how to create output fast (no distinct, no order) and how to have distinct and/or order
• [ ] cleanup of releases with broken jars
• [x] documentation of which add-opens flags need to be passed for spark: https://github.com/SmartDataAnalytics/RdfProcessingToolkit/blob/jena-5.0.0/rdf-processing-toolkit-pkg-parent/rdf-processing-toolkit-pkg-deb-cli/src/deb/resources/usr/bin/rpt#L12
• [x] Mapping files to graphs with rpt integrate graph=urn:foo file1 file2 graph=urn:bar file3 file4
• [x] There should be dedicated sections for how to use rpt in different architecture setups; e.g. as a caching proxy.

Thanks to @SimonBin the maven build produces ascii documentation files for the picocli commands, but they need to be added to the ...

• [ ] documentation web page.
• [ ] deb / rpm packaging