Context-Enhanced Multi-View Trajectory Representation Learning: Bridging the Gap through Self-Supervised Models