Dataset of RCQE (Automatic Code Quality Estimation in multi-round Code Review)

A multi-modal dataset , including source code, code diffs, and natural language comments as three types of metadata. The format for each instance in the dataset is as follows: code source file (in source code format), current commit code (in diff format), current commit comment (in natural language format, empty if none), previous commit (in diff format), previous submission commit comment (in natural language format).