I tried to have a look here, but running the script with the 5 maps takes much too long. I can't do a test that takes hours, just to run it for hours again later on to see if it is fixed. Test cases need to be small and short.
I did place points with a bigger spacing and in that case I see randomness with 1 map and also with 5 maps. The last one seems more dense, almost as if the different map files contain the same features, so they are processed multiple times.
when I read toc files, it's longitude element that make series with the same number. Randomness works fine for longitude and lattitude element at beginning of process but sometime randomness for longitude stops and make series.