Measuring Large Language Models Capacity to Annotate Journalistic Sourcing