feat: Add Infrahub Jinja2 transform for VRF/L3VNI configuration (#21)

- Add GraphQL query vrf_intent.gql: queries InfraVRFDeviceAssignment by
  device name, returning device-specific RD, VRF attributes (name,
  description, l3vni), and both device-level and VRF-level import/export
  route targets for fallback logic
- Add Jinja2 template vrf_yang.j2: renders a JSON array of VRF config
  objects with RD fallback (device → VRF-level) and RT merge logic
  (device-level if present, otherwise VRF-level); all optional
  relationships guarded with 'is defined and is not none' checks
- Update .infrahub.yml: register vrf_intent query and vrf_yang_transform
- Add unit test fixtures for leaf3: input.json (mock GraphQL response)
  and output.json (expected rendered JSON array)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

.infrahub.yml

@@ -24,6 +24,8 @@ queries:
     file_path: infrahub/transforms/queries/interface_intent.gql
   - name: vxlan_intent
     file_path: infrahub/transforms/queries/vxlan_intent.gql
+  - name: vrf_intent
+    file_path: infrahub/transforms/queries/vrf_intent.gql
 
 jinja2_transforms:
   - name: vlan_yang_transform
@@ -38,3 +40,7 @@ jinja2_transforms:
     description: "Generate VXLAN/VTEP configuration payload from Infrahub intent"
     query: vxlan_intent
     template_path: infrahub/transforms/templates/vxlan_yang.j2
+  - name: vrf_yang_transform
+    description: "Generate VRF/L3VNI configuration payload from Infrahub intent"
+    query: vrf_intent
+    template_path: infrahub/transforms/templates/vrf_yang.j2