Finish running optimizations and predictions for unfinished compounds

[hechth]

• Collect list of compounds which did not converge in optimization
• Check if the non-coverging molecules have acceptable geometry and filter out the ones that don't
• Run with normal spectra prediction metacentrum workflow
• Run spectra predictions
• Add newly predicted spectra into the repository
• Find one missing molecule
• Open up pull request which contains the data which was newly predicted

Optional:

• Try optimizing those molecules using XTB on Galaxy

[hechth]

All molecules inspected seem fine and only 2 have a complex structure, so it is a bit unclear why they failed. We will proceed with optimizing them with XTB2 and then continue with the predition.

[hechth]

Almost all molecules are done and MSP is collected, only one spectrum is missing - we need to figure out what happened here.

[hechth]

@wverastegui please open a PR to add those predicted spectra here to the repository :)