Sotorrent Data Set

Stack Overflow (SO) is the largest Q&amp;A website for software developers, providing a huge amount of copyable code snippets. Recent studies have shown that developers regularly copy those snippets into their software projects, often without the required attribution. Beside possible licensing issues, maintenance issues may arise, because the snippets evolve on SO, but the developers who copied the code are not aware of these changes. To help researchers investigate the evolution of code snippets on SO and their relation to other platforms like GitHub, we build <em>SOTorrent</em>, an open data set based on data from the official SO data dump and the Google BigQuery GitHub data set. <em>SOTorrent</em> provides access to the version history of SO content on the level of whole posts and individual text or code blocks. Moreover, it links SO content to external resources in two ways: (1) by extracting linked URLs from text blocks of SO posts and (2) by providing a table with links to SO posts found in the source code of all projects in the BigQuery GitHub data set.堆栈溢出（Stack Overflow，简称SO）是面向软件开发者的规模最大的问答网站，坐拥海量可直接复用的代码片段。近期研究显示，开发者常会将这些代码片段直接复制到自身的软件项目中，且往往未标注必要的归属信息。除可能引发授权许可纠纷外，还可能产生维护层面的问题：由于代码片段在SO平台上持续迭代演化，但复制代码的开发者并未知晓这些更新变更。为助力研究者探究SO平台上代码片段的演化规律，以及其与GitHub等其他平台的关联，我们基于SO官方数据导出集与谷歌BigQuery GitHub数据集，构建了开源数据集SOTorrent。SOTorrent支持访问SO内容的版本历史，支持的粒度覆盖完整帖子、独立文本块与代码块。此外，该数据集通过两种方式将SO内容与外部资源建立关联：（1）从SO帖子的文本块中提取其所链接的统一资源定位符（URL）；（2）提供关联表，收录谷歌BigQuery GitHub数据集中所有项目的源代码中出现的SO帖子链接。