Python scrapy when processing process_item, how can we process previous objects?

Question

When using Python scrapy and more particular, when processing process_item in pipelines. Then how can we process previous objects?

For example. We loop through class