bohr/bohr.lock

{
    "config": {
        "bohr_framework_version": "0.4.10",
        "tasks": {
            "bugginess": {
                "description": "",
                "top_artifact": "bohr.collection.artifacts.commit.Commit",
                "label_categories": [
                    "CommitLabel.NonBugFix",
                    "CommitLabel.BugFix"
                ],
                "test_datasets": [
                    "1151-commits",
                    "berger",
                    "developer-labeled-commits",
                    "herzig"
                ],
                "train_datasets": [
                    "200k-commits",
                    "fine-grained-refactorings"
                ],
                "label_column_name": "bug"
            },
            "smells": {
                "description": "",
                "top_artifact": "bohr.collection.artifacts.method.Method",
                "label_categories": [
                    "SnippetLabel.NonSmelly",
                    "SnippetLabel.Smelly"
                ],
                "test_datasets": [
                    "smells-test"
                ],
                "train_datasets": [
                    "smells-train"
                ],
                "label_column_name": "smelly"
            }
        },
        "datasets": {
            "berger": {
                "author": "Berger, E. D., Hollenbeck, C., Maj, P., Vitek, O., & Vitek, J.",
                "description": "Dataset of labeled commits from Berger, E. D., Hollenbeck, C., Maj, P., Vitek, O., & Vitek, J. (2019). On the impact of programming languages on code quality: a reproduction study. ACM Transactions on Programming Languages and Systems (TOPLAS), 41(4), 1-24.",
                "path": "berger.csv",
                "path_preprocessed": "berger.csv",
                "preprocessor": "copy",
                "loader": "csv",
                "mapper": "bohr.collection.datamappers.commit.CommitMapper"
            },
            "herzig": {
                "author": "K Herzig, S Just, A Zeller",
                "description": "Dataset of manually classified bug reports from K Herzig, S Just, A Zeller It's not a bug, it's a feature: how misclassification impacts bug prediction with linked commits recovered by giganticode",
                "path": "herzig.csv",
                "path_preprocessed": "herzig.csv",
                "preprocessor": "copy",
                "loader": "csv",
                "mapper": "bohr.collection.datamappers.commit.CommitMapper"
            },
            "1151-commits": {
                "author": "Levin, S., Yehudai, A.",
                "description": "Dataset of manually-labeled 1k commits (corrective, perfective, adaptive) from Levin, S., & Yehudai, A. (2017, November). Boosting automatic commit classification into maintenance activities by utilizing source code changes. In Proceedings of the 13th International Conference on Predictive Models and Data Analytics in Software Engineering (pp. 97-106).",
                "path": "1151-commits.csv",
                "path_preprocessed": "1151-commits.csv",
                "preprocessor": "copy",
                "loader": "csv",
                "mapper": "bohr.collection.datamappers.commit.CommitMapper"
            },
            "200k-commits": {
                "author": "JA. Prenner",
                "description": "Dataset of commits mined by JA. Prenner from Github.\n Cite: @inproceedings{babii2021mining,\n  title={Mining Software Repositories with a Collaborative Heuristic Repository},\n  author={Babii, Hlib and Prenner, Julian Aron and Stricker, Laurin and Karmakar, Anjan and Janes, Andrea and Robbes, Romain},\n  booktitle={2021 IEEE/ACM 43rd International Conference on Software Engineering: New Ideas and Emerging Results (ICSE-NIER)},\n  pages={106--110},\n  year={2021},\n  organization={IEEE}\n}. License: Creative Commons Attribution 4.0 International",
                "path": "200k-commits.csv",
                "path_preprocessed": "200k-commits.csv",
                "preprocessor": "copy",
                "loader": "csv",
                "mapper": "bohr.collection.datamappers.commit.CommitMapper"
            },
            "200k-commits-issues": {
                "author": "JA. Prenner",
                "description": "200k-commit dataset: linked issues. License: Creative Commons Attribution 4.0 International",
                "path": "200k-commits-issues.csv",
                "path_preprocessed": "200k-commits-issues.csv",
                "preprocessor": "copy",
                "loader": "csv",
                "mapper": "bohr.collection.datamappers.issue.IssueMapper",
                "dtype": {
                    "labels": "str"
                },
                "keep_default_na": false
            },
            "200k-commits-files": {
                "author": "JA. Prenner",
                "description": "200k-commit dataset: linked files. License: Creative Commons Attribution 4.0 International",
                "path": "200k-commits-files.csv.7z",
                "path_preprocessed": "200k-commits-files.csv",
                "preprocessor": "7z",
                "loader": "csv",
                "mapper": "bohr.collection.datamappers.commitfile.CommitFileMapper"
            },
            "200k-commits-link-issues": {
                "author": "JA. Prenner",
                "description": "Dataset of links between 200k-commits and 200-commits issues. License: Creative Commons Attribution 4.0 International",
                "path": "200k-commits-link-issues.csv",
                "path_preprocessed": "200k-commits-link-issues.csv",
                "preprocessor": "copy",
                "loader": "csv"
            },
            "200k-commits-manual-labels": {
                "author": "All contributors: H. Babii",
                "description": "Manual labels for the 200k-commits dataset. License: Creative Commons Attribution 4.0 International",
                "path": "200k-commits-manual-labels.csv",
                "path_preprocessed": "200k-commits-manual-labels.csv",
                "preprocessor": "copy",
                "loader": "csv",
                "mapper": "bohr.collection.datamappers.manuallabels.ManualLabelMapper"
            },
            "smells-train": {
                "author": "Madeyski, Lech, Lewowski, Tomasz",
                "description": "Dataset from Madeyski, Lech, and Tomasz Lewowski. \"MLCQ: Industry-relevant code smell data set.\" preprocessed by H. Babii to extract long method smells and split into train and test sets. License: Creative Commons Attribution 4.0 International",
                "path": "smells-madeyski.csv",
                "path_preprocessed": "smells/train.csv",
                "preprocessor": "data-preprocessing/smells.sh",
                "loader": "csv",
                "mapper": "bohr.collection.datamappers.method.MethodMapper",
                "sep": ";"
            },
            "developer-labeled-commits": {
                "author": "Andreas Mauczka, Florian Brosch, Christian Schanes, Thomas Grechenig",
                "description": "Dataset of Developer-Labeled Commit Messages",
                "path": "developer-labeled-commits.zip",
                "path_preprocessed": "developer-labeled.csv",
                "preprocessor": "data-preprocessing/developer_labeled.py",
                "loader": "csv",
                "mapper": "bohr.collection.datamappers.commit.CommitMapper"
            },
            "fine-grained-refactorings": {
                "author": "Krasniqi, Rrezarta; Cleland-Huang, Jane",
                "description": "dataset of manually-labeled fine-grained refactoring types from Enhancing Code Refactoring Detection with Explanations from Commit Messages",
                "path": "fine-grained-refactorings.zip",
                "path_preprocessed": "fine-grained-refactorings.csv",
                "preprocessor": "data-preprocessing/fine-grained-refactorings.py",
                "loader": "csv",
                "mapper": "bohr.collection.datamappers.commit.CommitMapper"
            },
            "smells-test": {
                "author": "Madeyski, Lech, Lewowski, Tomasz",
                "description": "Dataset from Madeyski, Lech, and Tomasz Lewowski. \"MLCQ: Industry-relevant code smell data set.\" preprocessed by H. Babii to extract long method smells and split into train and test sets. License: Creative Commons Attribution 4.0 International",
                "path": "smells-madeyski.csv",
                "path_preprocessed": "smells/test.csv",
                "preprocessor": "data-preprocessing/smells.sh",
                "loader": "csv",
                "mapper": "bohr.collection.datamappers.method.MethodMapper",
                "sep": ";"
            }
        },
        "dataset-linkers": [
            {
                "from": "200k-commits",
                "to": "200k-commits-files"
            },
            {
                "from": "200k-commits",
                "to": "200k-commits-issues",
                "link": "200k-commits-link-issues"
            },
            {
                "from": "200k-commits",
                "to": "200k-commits-manual-labels"
            }
        ]
    },
    "heuristics": {
        "heuristics/bugginess.py": "a55520b99cbaad572a26886b52e74652",
        "heuristics/manuallabels.py": "f338b2a285d76da97b3f53e9b167368a",
        "heuristics/smells.py": "b1a5ed3a14eb9eae8924b8a43e3bc452"
    },
    "manual_stages": {}
}