This repository provides a comprehensive framework for mining, collecting, analyzing, and modeling code merge scenarios, with a focus on conflict resolution and edit script inference. The project integrates data mining, dataset creation, model training/evaluation, and API serving, supporting both research and practical applications.

• EditFusion_implementation/

  • Main implementation for model training, evaluation, and inference.
  • Includes a Flask-based REST API for serving predictions.
  • Contains scripts for dataset collection, model definition, training, evaluation, and inference.
  • See EditFusion_implementation/README.md for detailed usage and setup.

• gitMergeScenario/

  • Java-based tools for mining and collecting conflict files and tuples from Git repositories.
  • Supports automated and manual repository analysis, conflict tuple extraction, and dataset generation.
  • See gitMergeScenario/README.md for setup, input/output formats, and troubleshooting.

• GraphQL_repo_mining/

  • Python scripts and notebooks for mining GitHub repositories using the GraphQL API.
  • Enables filtering, analysis, and extraction of repository metadata based on custom criteria.
  • See GraphQL_repo_mining/README.md for mining instructions and requirements.

• data_collect_analysis/

  • Scripts and notebooks for collecting, analyzing, and evaluating code merge algorithms and datasets (e.g., MergeBert).
  • Supports experiment automation, data collection from popular repositories, and algorithm evaluation.
  • See data_collect_analysis/README.md for experiment setup and usage.

1. Mining and Collecting Data

   • Use gitMergeScenario to collect conflict files and tuples from repositories.
   • Use GraphQL_repo_mining to mine repository metadata from GitHub.
   • Use data_collect_analysis for additional data collection and algorithm evaluation.

2. Dataset Preparation

   • Aggregate and preprocess conflict data using scripts in EditFusion_implementation/train_and_infer/.
   • See the respective README files for data format and preprocessing steps.

3. Model Training and Evaluation

   • Train and evaluate models using scripts in EditFusion_implementation/train_and_infer/.
   • Adjust dataset/model paths as needed.
   • See EditFusion_implementation/README.md for details.

4. API Service

   • Serve model predictions via the Flask API in EditFusion_implementation/flask_service/.
   • See EditFusion_implementation/README.md for API usage and endpoints.

• Python 3.x (for Python scripts)
• Java (for gitMergeScenario)
• See each subdirectory's README for specific package and environment requirements.

• Each subdirectory contains a dedicated README with detailed instructions, usage examples, and troubleshooting tips.
• For advanced usage, dataset formats, and experiment configurations, refer to the README in the relevant directory:
  • EditFusion_implementation/README.md
  • gitMergeScenario/README.md
  • GraphQL_repo_mining/README.md
  • data_collect_analysis/README.md

For questions, issues, or contributions, please open an issue or pull request in this repository.