Fix pg-backup-api so it can handle Barman <= 3.9

pg-backup-api 2.1.0 doesn't work with Barman <= 3.9 because it attempts to
fetch models independently of the Barman version it operates on. As Barman
models were introduced through version 3.10, pg-backup-api fails with an
ugly stack trace:

```
Traceback (most recent call last):
  File "/usr/lib/python3.9/site-packages/flask/app.py", line 2073, in wsgi_app
    response = self.full_dispatch_request()
  File "/usr/lib/python3.9/site-packages/flask/app.py", line 1518, in full_dispatch_request
    rv = self.handle_user_exception(e)
  File "/usr/lib/python3.9/site-packages/flask/app.py", line 1516, in full_dispatch_request
    rv = self.dispatch_request()
  File "/usr/lib/python3.9/site-packages/flask/app.py", line 1502, in dispatch_request
    return self.ensure_sync(self.view_functions[rule.endpoint])(**req.view_args)
  File "/usr/lib/python3.9/site-packages/pg_backup_api/logic/utility_controller.py", line 76, in diagnose
    available_models = barman.__config__.model_names()
AttributeError: 'Config' object has no attribute 'model_names'
```

Through this commit we fix pg-backup-api in that sense:

* It attempts to use models and the new signature of the `exec_diagnose`,
  which were introduced by Barman 3.10;
* It falls back to the old signature of `exec_diagnose` if models are not
  available (Barman <= 3.9).

References: BAR-171.

Signed-off-by: Israel Barth Rubio <[email protected]>

pg_backup_api/pg_backup_api/logic/utility_controller.py

@@ -73,18 +73,24 @@ def diagnose() -> 'Response':
         else:
             server_dict[server] = Server(conf)
 
-    available_models = barman.__config__.model_names()
-    model_dict = {}
-    for model in available_models:  # pyright: ignore
-        model_dict[model] = barman.__config__.get_model(model)
-
     # errors list with duplicate paths between servers
     errors_list = barman.__config__.servers_msg_list
 
-    barman_diagnose.exec_diagnose(server_dict,
-                                  model_dict,
-                                  errors_list,
-                                  show_config_source=False)
+    try:
+        available_models = barman.__config__.model_names()
+        model_dict = {}
+        for model in available_models:  # pyright: ignore
+            model_dict[model] = barman.__config__.get_model(model)
+
+        barman_diagnose.exec_diagnose(server_dict,
+                                      model_dict,
+                                      errors_list,
+                                      show_config_source=False)
+    # An attribute error is thown when calling `model_names()` if using Barman
+    # older than 3.10, in which case models are not yet implemented, so we fall
+    # back to the old signature of diagnose command.
+    except AttributeError:
+        barman_diagnose.exec_diagnose(server_dict, errors_list)
 
     # new outputs are appended, so grab the last one
     stored_output = json.loads(output._writer.json_output["_INFO"][-1])