fix: skip header row in retire_user CSV

[rahulkanneri-2u]

Description

Updates the retire_user management command to skip a CSV header row when the first two columns are username,email or username,user_email.

This fixes a Stage retire_users_from_csv failure where the command treated the header row as a real user and failed with:

Could not find users with specified username and email address: [{'username': 'email'}]

Impact: Operators running the retirement management command with a CSV file can now use a CSV that includes a standard header row. No learner-facing UI impact.

Supporting information

Private Jira: BOMS-717

Testing instructions

1. Create a CSV with a header row and one valid user:

   username,email
   <username>,<email>

2. Run:

   python manage.py lms retire_user --user_file <path-to-file>

3. Confirm the command skips the header row and processes the real user.

Automated coverage added:

• Added a regression test for retire_user --user_file with a username,email header row.

Local verification:

• Ran python -m py_compile on the touched command and test files.
• Targeted pytest could not run locally because this environment is missing kombu during conftest.py import.

Deadline

None

Other information

This does not include the edx-internal cleanup that will eventually re-enable notifications. That can happen after this fix is deployed to Stage and the retire_users_from_csv job is confirmed passing.

No migrations, UI changes, or accessibility concerns.