why the dataset and the output of the env(action) has the element of "discount"but not the "done"?

Hi, when I use the dataset and the env of the UnpluggedRL , I met some things that confused me.First , When I downloaded the training data of the cheetah_run, one of the dm_control_suite,while,I just got the tuple (obs, nest_obs,action, next_action, reward, discount), which doesn't include the "done" flag, so as to the output of the env(action). Now I don't know how to use the "discount", is it equal to "done"?
I have used the discount element as the done, but when I trained the policy network by BC,the result is very bad,it seems the network learns nothing, the training result of the other algorithms are also very bad.What should I do,would you have some suggestions?