“…Owing to the large sample size, this study employed Python scripts (3.8.5) to read those judgement files with PyCharm (Edition 2020.2.1). To replicate as closely as possible the mechanism of how humans read, to re-articulate the content in the script, and deal with the format in which judgement files are normally recorded, the script to identify certain sections of the files (including the introduction, evidence, sentencing) was written based on the common framework and expressions frequently used by and taught to practitioners [34]. Although the manual by Hu [34] suggested many formal phrases to introduce each section of a judgement file, there were many variations in the actual wording.…”