Query regrading quantitative results

Hi,

I am interested to know how the quantitative results like Precision and F! score were generated for a synthetic dataset. I tried to read more details in paper but I could not find much.

It would great if you could share how to generate those :))