My idea was to pass all above including main.bash as parameters to the
python wrapper (called run.py, attached), as follows:

where the last parameter sys.argv[7] is the output file, specified by xml file.

This is the correct approach.

But galaxy was running for more than 3 hours, although it should take
only 7-10 minutes on our server with no results.

Vispa should write its output to dataset_192_I_2_20_CNTGS_DIST0_EM20.txt
, where dataset_192 .txt is the reads file.But dataset_192_I_2_20_CNTGS_DIST0_EM20.txt did not appear in galaxy-dist/database/files/000.
run.py: simply coppies data from dataset_192_I_2_20_CNTGS_DIST0_EM20.txt to galaxy output file, but nothing is coppied since dataset_192_I_2_20_CNTGS_DIST0_EM20.txt was not created.

How could the results appear?Could somebody help me ?

The first thing to check is if run.py functions correctly outside of Galaxy. If so, try checking the job manually using top/qstat/similar system utilities to figure out where the job is stuck.