Clarify the input handling of session template CRUD API

[achimnol]

In lablup/backend.ai-manager#480, #1406, and #1668, the session template creation API became able to process only the yaml.safe_load_all()'s return value, which is a generator.

backend.ai/src/ai/backend/manager/api/session_template.py

Lines 57 to 65 in 3968538

	try:
	body = json.loads(params["payload"])
	except json.JSONDecodeError:
	try:
	body = yaml.safe_load_all(params["payload"])
	except (yaml.YAMLError, yaml.MarkedYAMLError):
	raise InvalidAPIParameters("Malformed payload")
	for st in body:
	template_data = check_task_template(st["template"])

In L64, it may crash with the json.loads() output as it would just return a dict, when the client submits a JSON data instead of YAML data. This is a potential bug.

Along with #3183, we need to clarify whether:

• the API handler decomposes the input parameter object and applies individual validation to its parts, or
• delegates the entire parameter interpretation to a structured schema (e.g., Pydantic).

The template schema itself should also allow unspecified value for the architecture field to support reusing the same template with multi-arch images in different host architectures (e.g., DGX-H100 and GH200 nodes).

cf) Currently the manager accepts as payload of the session template creation API:

template:
  api_version: 1  # actually ignored
  type: "taskTemplate"
  spec:
    kernel:
      ...