Data for Multilingual logs parsing

This is the dataset and intermediate data for paper: Beyond English Logs: An Empirical Study of Parsing Logs with Not(-only)-English Contents Translated_log_generation.zip contains the original English log datasets and the translated multilingual datasets: original folder contains the original English datasets that contain groundtruth and Origin folder contains the original log files only. translated_log contains the translated template file, translated log files and the groundtruth event files for multilingual datasets that are generated by our approach translated_log_baseline contains the translated multilingual datasets that are generated by directly translated baseline. Userstudy contains the files for user study to evaluate the accuracy of translated logs Parsing_tools_evaluation.zip contains the files for evaluating the performance of existing log parsing techniques on multilingual logs: group_check_result contains the group accuracy checking result (groundtruth) the group consistency checking result (ori_parsing_result) parsing_result contains the parsing result of English datasets and mulitilingual datasets tokenized_log contains the tokenized result for multilingual datasets downstream_data.zip contains the files for DeepLog downstream task evaluation, each folder contains: 6 folders contains data for English, Chinese, French, Japanese, Korean, and Mix language files For each language, the folder contains: log_structured.csv file contains full parsing result train.log_structured.csv file contains parsing result used for training test.log_structured.csv file contains parsing result used for test train and test file containing feature vectors generated by preprocess.py