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