Invert veto file usage in PyGRB

[pannarale]

The PyGRB workflow generator fills up the veto-file with science segments that are okay to be used (e.g., times that are not CAT2 vetoed) rather than problematic. The post-processing scripts correctly use this file. The problem uncovered by PRs #5038 and #5042 is that if there are no problematic (e.g., CAT2) times in the stretch of date being analysed, the veto-file is empty.

The clean thing to do is to:

1. -CAT2 in the configuration file, rather than +CAT2 (for example);
2. invert the logic in the workflow generator here (this should directly grab vetoed_segs;
3. also invert the logic in this utility which should again directly grab vetoed segments; note that this affects this function which ends up refering to good segments as vetoes;
4. finally, swap the calls to veto.indices_outside_segments and veto.indices_within_segments in this function which called by all postprocessing scripts that are meant to apply additional dq vetoes to injections/triggers.