[BUG] Upgrade doesn't work with ERROR epicli - No such attribute: cloud

[atsikham]

Describe the bug
Upgrade process finishes with an error

ERROR epicli - No such attribute: cloud

To Reproduce
Steps to reproduce the behavior:

1. Deploy v0.6 cluster on Azure.
2. Run upgrade from develop branch.

Expected behavior
Successful upgrade.

Config files

---
kind: configuration/shared-config
title: Shared configuration that will be visible to all roles
name: default
provider: azure
specification:
  custom_repository_url: ''
  custom_image_registry_address: ''
  download_directory: /tmp
  vault_location: ''
  vault_tmp_file_location: SET_BY_AUTOMATION
  use_ha_control_plane: true
  promote_to_ha: false
version: 0.7.1
---
kind: epiphany-cluster
title: Epiphany cluster Config
provider: azure
name: odfe
specification:
  prefix: atsikham
  name: certs
  admin_user:
    name: ubuntu
    key_path: /home/vscode/.ssh/id_rsa
  cloud:
    subscription_name: <subscription>
    use_public_ips: true
    use_service_principal: false
    region: West Europe
  components:
    kubernetes_master:
      count: 1
    kubernetes_node:
      count: 0
    logging:
      count: 0
    monitoring:
      count: 0
    kafka:
      count: 0
    postgresql:
      count: 0
    load_balancer:
      count: 0
    rabbitmq:
      count: 0
    ignite:
      count: 0
    opendistro_for_elasticsearch:
      count: 1
    single_machine:
      count: 0
version: 0.7.1

OS (please complete the following information):

• OS: Ubuntu, but it doesn't matter

Cloud Environment (please complete the following information):

• Cloud Provider: Azure, but all affected

Additional context
There is such piece of code in AnsibleInventoryUpgrade.py:

# re-constructure cluster model with all data necessary to run required upgrade rolls
        self.cluster_model = dict_to_objdict({
            'provider': 'any',
            'specification': {
                'admin_user': {
                    'name': loaded_inventory.groups['all'].vars['ansible_user'],
                    'key_path': loaded_inventory.groups['all'].vars['ansible_ssh_private_key_file']
                }
            }
        })

That model is used then in upgrade process with no success, see related changes.

[to-bar]

It looks like a case similar to #909

[atsikham]

will be tested by @to-bar together with #1647

[to-bar]

Test OK (from 0.7.2 to develop).