added censor variable + extracted km cohort

[marrpesce]

To create the KM cohort:

• I added to the dataset_definition_snapshot.py the first deregistration date after the snapshot date
• I added to the data_combined table
  • the death_date and dereg_date to build the censor variable
    • I can combine them on a collective censor variable if needed
  • the start_date (=snapshot_date)and end_date (= next_campaign_date)
• I updated the report_snapshot to save the data_combined table as km_cohort.csv
  • Should we drop the ID variable?
• I didn't modify the project.yaml file, but we should modify the output to highly_sensitive since it includes the km_cohort.csv