Code diffs and commit messages from top1000-2000 Java projects in GitHub

This dataset contains pairs <code changes, commit message> from top 1000-2000 Java projects in GitHub via GitHub Developer API.

The structure of files is as follows:

• <repository-id>.<repository-name>/       # Each directory contains commits of one project.
  • <commit-id>.<commit-sha>/          # Each directory contains one commit information, and
                                                       # the commit-sha is the primary key of this commit in GitHub. 
                                                       # The children in this directory have 3 types: the file
                                                       # "commit_msg", multiple directories "<filename>", and the file
                                                       # "error". 
    • commit_msg                           # This file contains one line, representing the commit
                                                   # message. 
    • <filename>/                            # The name of this directory is the changed file name.
      • patch                            # The code diffs of this <filename>, describing
                                            # which lines are added and which lines are
                                            # removed with some same lines context.
    • <filename>/
      • patch
    • error                                       # This file contains the error message when crawling
                                                    # from GitHub. If this file exists, the directories
                                                    # <filename> will not exist.