From Street View Imagery to Street Quality Indicators: Vision Language Inference for the Suburban 15-minute City