Recreate path selection

[smnnlt]

I do not really understand the current process of setting the data and tools paths, where you have to write the paths in a .txt file of a particular structure which is then imported into Matlab to extract the pathes as variables. In my view, it would be much easier to require the user to set the paths as a variable at the beginning of the script (or even better as function arguments).

Maybe it is also possible to auto-detect the tools paths with which. This would make make it easier to get the software started as only the data path has to be given. However, you should then implement the option to manually select the tool paths as a fall back if the auto-detection fails.

Comment for the JOSS review (openjournals/joss-reviews#7858).

[sfregosi]

I also had some trouble navigating the existing path set up process. I've added my comments below:

The current path set up with paths.txt is not super robust to mistakes

• The paths.txt I downloaded had "## to start the header lines (different than wiki screenshot, also different than standard MATLAB comment line). Either make wiki and paths.txt consistent, or better suggest changing to % to match MATLAB. Also add a line to the paths.txt that specifies how to mark a 'comment' line.
• Reading in the path values of paths.txt relies on exact spacing and number of header lines. I initially just commented out your default paths so I'd have a reference to them, so I had extra header lines and the tools, mat, and data paths were not read in correctly in main.m lines 76-78. Alternatively, you could use eval (although use of eval is controversial, I think this is a place in which it might make sense) and parse in these paths directly in the paths.txt file. I think this would be more flexible to differing spacing and number of lines.

My suggested improvements above would not be necessary by implementing #1. I wholly agree with @smnnlt 's suggestion to have these path as function input arguments.