Fix bypass pair validation to allow default bypass pair conflicts

[hildebrau]

Fix bypass pair validation to allow default bypass pair conflicts

Fixes #10

• Modified get_parent_child_dict() to distinguish between real port conflicts and false positives from default bypass pairs
• Allows bypass pairs to use ports already in used_parent_name_list if those ports belong to other bypass pairs (which will be removed)
• Fixes round-trip workflow (pull_site -> do_site) broken in SDK 6.6.1b1
• Resolves issue where user-specified bypass pairs were incorrectly flagged as conflicts with default bypass pairs

Description

This PR fixes a validation error in get_parent_child_dict() that incorrectly flags user-specified bypass pairs as conflicts when they use ports that are already in used_parent_name_list from default bypass pairs.

The Problem:
When processing config_interfaces_defaults (which includes both default model interfaces and user-specified interfaces), default bypass pairs are processed first and add their WAN/LAN ports to used_parent_name_list. When user-specified bypass pairs are processed later, the validation incorrectly flags them as conflicts, even though the default bypass pairs will be removed (as they should be when user specifies a config).

The Solution:
Modified the validation logic to check if a port conflict is with another bypass pair before throwing an error. This allows bypass pairs to use ports that are already in used_parent_name_list if those ports belong to other bypass pairs, while still preventing real conflicts with subinterfaces, PPPoE interfaces, virtual interfaces, etc.

Code Changes:

• Modified prisma_config/do.py lines 5570-5603 in function get_parent_child_dict()
• Added logic to distinguish between bypass pair conflicts (allowed) and non-bypass pair conflicts (real errors)

Motivation and Context

This PR addresses issue #10. After upgrading to prisma_sase 6.6.1b1 with prisma_config 6.5.1b2, the round-trip workflow (pull_site → do_site) fails with:

ERROR: Bypass pair bp_45 is using a port that is a parent of another interface

This breaks critical workflows:

• Configuration backup and restore
• Configuration version control
• CI/CD pipelines that validate configurations
• Testing configuration changes

The configuration is valid in the controller (it was exported from there), but do_site rejects it during validation. This is a regression that prevents users from performing basic configuration management tasks.

How Has This Been Tested?

Test Environment:

• prisma_config version: 6.5.1b2
• prisma_sase SDK version: 6.6.1b1
• Python version: 3.9
• Operating System: Linux (RHEL 9.6)
• Test site: ITLAB1 (contains bypass pairs with servicelink interfaces)

Test Cases:

1. ✅ Existing YAML file: Ran do_site on an existing YAML file that previously failed - SUCCESS
2. ✅ Round-trip test: pull_site -S ITLAB1 --output ITLAB1_test.yml followed by do_site ITLAB1_test.yml - SUCCESS
3. ✅ Configuration application: Verified configuration applied correctly to both elements in the site

Testing Limitations:

• Testing was performed on a single site (ITLAB1) with bypass pairs and servicelink interfaces
• The fix was validated for the specific error scenario described in issue Bypass Pair Validation Error After SDK Upgrade #10
• Limited regression testing was performed - only confirmed that the specific bypass pair validation error was resolved and that the site configuration applied successfully
• Additional testing across different site configurations, interface types, and edge cases would be beneficial

Test Output:

No Change for Site ITLAB1.
...
Updated Interface bp_45(45) (Etag 7 -> 8).
...
DONE

The fix successfully allows bypass pairs to be processed without false validation errors while still preventing real conflicts.

Screenshots (if appropriate)

N/A - Terminal output provided in testing section above.

Types of changes

• Bug fix (non-breaking change which fixes an issue)

Checklist

• I have updated the documentation accordingly. (N/A - no documentation changes needed)
• I have read the CONTRIBUTING document. (Will review before submitting)
• I have added tests to cover my changes if appropriate. (N/A - manual testing performed, no test framework in place)
• All new and existing tests passed. (Manual testing on specific use case confirms fix works - see Testing Limitations above)