every job has a name that reflects the Clickhouse database used to backup the results. Just passing this job_name should allow continuing from a checkpoint.
when restarting a stopped job, we should perform a step-wise verification to determine the latest step that is entirely complete and valid.
any new_step call should be fully reproducible later because of the completion_step and inactivation_step columns. Implement and test this.
Write a separate continue(...) function that restarts a job. Ideally this only requires the job_name as parameter.
completion_stepandinactivation_stepcolumns. Implement and test this.continue(...)function that restarts a job. Ideally this only requires the job_name as parameter.