---

This repository contains the code and data for paper ‘Quantifying Community Evolution in Developer Social Networks’.

Requirements

You need Python 3.9, R, and RStudio to run the code.

Important files

• The entry of Python code for community detection and index calculation is in main_calculate_indexes.py
• Data are located at data/
• After executing main_calculate_indexes.py, the results will be recorded in an auto-generated subfolder in the result/ folder of the project
• The R script RScript/productivity_analysis.Rmd is used to analyze the correlation between community evolution indices and team productivity

How to obtain the results

Concurrent Validity: Community evolution pattern detection

• Set the threshold of community matching used for existing approach in global_settings.py, e.g., EVOLUTION_PATTERN_THRESHOLD = 0.3
• Run main_calculate_indexes.py
• Find the results in concurrent_validity.txt in the auto-generated subfolder under result/

Discriminant Validity: Spearman’s Correlations Coefficients between pairs of indices

• Run main_calculate_indexes.py （Optional if you’ve already run the code for concurrent validity）
• Find the results in discrimant_validity.txt in the auto-generated subfolder under result/

Regression Analysis: Correlation with team productivity

• Run main_calculate_indexes.py （Optional if you’ve already run the code for concurrent validity）
• Find the path of file index_productivity.csv in the auto-generated subfolder under result/
  • For example: ../result/2022-03-13T10-07-10Z_interval_7_days_x_12/index_productivity.csv
• Modify line 22 of RScript/productivity_analysis.Rmd to specify the path of the data file.
  • For example:

    table_data<-read.table("../result/2022-03-13T10-07-10Z_interval_7_days_x_12/index_productivity.csv", head=T, sep=',', stringsAsFactors = FALSE)

• Run RScript/productivity_analysis.Rmd in RStudio, and get the results.

部署脚本

针对每个main_xxx.py（目前已经实现了xxx=main_calculate_indexes），需要添加两个部署脚本：

• ./deploy_x_main_xxx.sh：需要复制./deploy_0_main_calculate_indexes.sh脚本，重命名该脚本，并修改其中的TASK_SCRIPT路径为下面的server_scripts/run_xxx.sh文件路径
• server_scripts/run_xxx.sh：需要复制server_scripts/run_main_calculate_indexes.sh脚本，重命名该脚本，并修改其中的py文件名

完成上述脚本添加后，本地用Linux Shell打开项目目录（Windows平台推荐安装WSL2），并执行下列语句：

dos2unix ./deploy_x_main_xxx.sh #只需要曾经执行过一次就行
./deploy_x_main_xxx.sh

该脚本会提示输入密码三次，其功能为将本地的代码清理后上传172.27.135.32服务器，登录用户wangliang并自动运行，忽略以下提示信息：

bash: cannot set terminal process group (1469755): Inappropriate ioctl for device
bash: no job control in this shell

上传服务器的代码位于服务器目录：~/workspace/deploy/oss_community_evolution_indexes/，会保存每次上传的代码副本，以时间区分

运行结果会保存在服务器目录：~/workspace/data/oss_community_evolution_indexes/result/，以时间区分

在使用ssh登录服务器后，可以使用htop命令查看当前程序是否启动（是否有用户wangliang的python3程序在执行），同时在~/workspace/deploy/oss_community_evolution_indexes/目录中使用tail output.log命令查看程序输出情况。