| # CodeReview: Intelligent Code Review System |
|
|
| π [Paper](https://huggingface.co/papers/xxxx.xxxxx) | π [Project Page](https://codereview.github.io/) | πΎ [Released Resources](https://huggingface.co/collections/toolevalxm/codereview-67a978e28fd926b56a4f55a2) | π¦ [Repo](https://github.com/xmhtoolathlon/Annoy-DataSync) |
|
|
| We release the raw data for our processed CodeReview-Synth dataset, derived from the original dataset from BigCode team. |
|
|
| The data format for each line in the `code_review_samples.jsonl` is as follows: |
| ``` |
| { |
| "code_snippet": <the original code to be reviewed>, |
| "language": <programming language>, |
| "review_comments": <structured code review feedback>, |
| "severity": <issue severity level>, |
| "category": <review category type>, |
| "source": <the source of the raw code files>, |
| "meta": <meta information about this sample> |
| } |
| ``` |
|
|
| Some samples may have incomplete review comments due to filtering constraints during processing. |
|
|
| *Note: Code review suggestions are generated through automated analysis and may require human verification for production use. |