Implementing filter by shift this way has one consequence of fetching employees with empty default shift if filter_by_shift is checked but shift isn't set. For other filters like company, department etc, if there's no value for filters all employees are fetched.

One more caveat is, because we're getting data from two different doctypes, get_list isn't enough to ensure all the filters are "and"ed e.g if someone wants to fetch employees from department A and shift B, it won't be ensured because shift assignment may have employees belonging to different department.

It's better to use query builder to select employees in this case. And while we're half way there, might as well use that to separate unmarked and half day marked employees too. I'll contribute to this. Looks good enough for now.

Thanks for your feedback! @asmitahase . The cases you mentioned are already covered in the current filter logic.

1. The employee_list is first filtered to get the unmarked attendance through the _get_unmarked_attendance function, which ensures that all necessary conditions (such as department, grades or other filters) are applied.
2. Using the same filtered unmarked_attendance, I'm further refining it by checking the shift assignments (both default shifts and shift assignments). This ensures that only employees assigned to the relevant shift are included.

Please correct me if I’m wrong in the current logic. Should I work on the requested changes if the filtering with shift logic is incorrect?