Overhaul the file handling
There should be the following concepts
mainfile, the main output file, as usual
codefiles, other files generated or inputted to the code
auxfiles, other files in directory (and subdirectories?) respecting some heuristics
allfiles, files in same directory (currently
allfiles should not be part of processing, but deduced by the API on request.
- "codefiles" based on filename regexps from parser
- allow to load more files in GUI
- include sub directories
- download all, dataset download, etc. use "allfiles"
heuristicts to detect meaningful