Replication package for: E4R-Reviewer: Effective and Explainable Automated Code Review via End-to-End Reasoning-Guided Alignment

Replication package for: E4R-Reviewer: Effective and Explainable Automated Code Review via End-to-End Reasoning-Guided Alignment, including source code and data. Directory structure: ├── code/ # Source code │ ├── lora_merge.py # Merge the LoRA adapter into the full model │ ├── pipeline.sh # End-to-end training pipeline script │ ├── train_grpo.py # GRPO training │ ├── train_sft.py # SFT training │ ├── util/ # Utilities │ │ ├── dataset_grpo.py # GRPO dataset loader │ │ ├── dataset_sft.py # SFT dataset loader │ │ ├── eval_cr.py # Automatic metrics for CR │ │ ├── eval_qe.py # Automatic metrics for QE │ │ ├── prompt.py # Prompts │ │ └── reward.py # Reward functions │ └── vllm_infer.py # Inference with vLLM ├── datasets/ # Datasets │ ├── test_qe_dataset.xlsx # QE test set │ ├── test_rcg_cr_dataset.jsonl # RCG and CR test set │ └── train_dataset.jsonl # Training data └── README.md