adjacent allowed VLANs not updating when separated by a comma

[Arapacana]

I have just realized that the VLAN discovery is not working fully -- it does populate VLAN ACLs but not fully in many cases. As one example, the interface on a Cisco 4900M switch is provisioned thusly:

interface GigabitEthernet2/18
 switchport trunk allowed vlan 1,423,437,452,453,475,488,496,497,511,513,545
 switchport trunk allowed vlan add 555,587,603,606,611,627,628,654,743,754,840

The natlas graph for this connection shows:

P:b'gi2/18'
C:b'gi0/25'
Native 1
Allowed P:1,423,437,452,475,488,496,511,513,545,555,587,603,606,611,627,654,743,754,840
Allowed C:All

You see that natlas missed VLANs 453, 497, and 628.

I believe I have pinpointed the issue to circumstances where a sequence of VLANs is entered into the ACL with commas, as in:

switchport trunk allowed vlan 452,453

vs using a dash:

switchport trunk allowed vlan 452-453

If a dash is used, natlas seems to pick up on the adjacent VLANs correctly, but if a comma is used to separate adjacent VLANs, it appears to fail, and only captures the first VLAN, from what I can tell.

Is there a way to fix the VLAN processing so that it accepts adjacent allowed VLANs when separated by a comma?

Thank you,
-Seth