PLIP: Language-Image Pre-training for Person Representation Learning