Code and Data for "PATIMT-Bench: A Multi-Scenario Benchmark for Position-Aware Text Image Machine Translation in Large Vision-Language Models" (EMNLP-25 findings) - View it on GitHub