A mixed-method in-depth study of test-specific refactorings: dataset
Bibliographic record
Abstract
This submission is the replication package and dataset of the paper "A mixed-method in-depth study of test-specific refactorings." The files are organized in folders according to the corresponding method. Method 1 (Mining datasets) has seven files, two compressed excel sheets and five archive files. Excel sheets: Victor Guerra Veloso - Review Matrix.xlsx.zstd: this spreadsheet comprises the data from the related work as it is exported by the scripts from Gerrit2CommitMapper and MethodOneRelatedFiles.tar.zstd. This is also the input file for the association rules experiment Validation Sample ItemsetsSupport AllLevels FullyMerged.xlsx.zstd: as the name suggests it's a sample of the fully merged output of the association rules experiment. Archives: Gerrit2CommitMapper.tar.zstd contains the Gerrit2CommitMapper script that we applied to Gerrit issues to extract associated commits in the projects' git repositories. MethodOneRelatedFiles.tar.zstd contains some scripts (as python notebooks) that were used to conduct the method one. related-work-datasets.tar.zstd contains all the raw datasets from related works MiningDatasetRMinerOutput.tar.zstd includes the output of RefactoringMiner for each commit in our dataset TestRefactoringExistingDataMining.tar.zstd comprises the python project we develop to assist the replication of the method 1's association rules experiment. It expects "Victor Guerra Veloso - Review Matrix.xlsx" as input and generate all the association rules for different settings, such as using different algorithms (FPGrowth, FPMax, and Apriori), hyper-parameters (minsupport ranging from 0.01 to 0.15), and transaction granularities (commit-level, file-level, and method-level) Method 2 (Monitoring) has two compressed text files, one excel sheet, and an archived version of the RefactoringMonitor tool source code. Notice that, RefactoringMonitor includes all components (WebApp, API, and worker), configuration files (docker-compose.yml), and a README.md with instruction for execution. Text files: InitialSet: list of repositories initially monitored (based on method 1) ExtendedSet: list of repositories monitored after collecting popular repositories (based on github search) Excel sheet: The excel sheet concentrates the result of the Monitoring and Survey as well as the multiple runs of the validation by the second author. Method 3 (StackOverflow) has one excel sheet and one archive file. The excel sheet contains the results of the experiment and associates tags to the collected and studied StackOverflow questions. The archive file comprise all the scripts to collect and assist the analysis of the stackoverflow questions. It also includes both input and output data files. These scripts leverage webbrowser, the Python's standard library feature to open browsers in a specific URL, to iterate over all stackoverflow questions allowing the researcher to assign tags or create new tags. Newly added files include 'Core Devs Reassessment.tar.zstd', 'MonitoringReposCached.tar.zstd', and 'MonitoringExtension.tar.zstd' which comprise the artifacts generated during the method 2 core devs reassessment, the archive of all method 2 monitored repos from which at least one refactoring type has been identified, and the the artifacts generated during the method 2 monitoring extension, respectively. List of files: [M1]Mining Datasets/Gerrit2CommitMapper.tar.zstd[M1]Mining Datasets/MethodOneRelatedFiles.tar.zstd[M1]Mining Datasets/MiningDatasetRMinerOutput.tar.zstd[M1]Mining Datasets/TestRefactoringExistingDataMining.tar.zstd[M1]Mining Datasets/Validation Sample ItemsetsSupport AllLevels FullyMerged.xlsx.zstd[M1]Mining Datasets/Victor Guerra Veloso - Review Matrix.xlsx.zstd[M2]Monitoring/Core Devs Reassessment.tar.zstd[M2]Monitoring/MonitoringReposCached.tar.zstd[M2]Monitoring/MonitoredRepositories-ExtendedSet.txt.zstd[M2]Monitoring/MonitoredRepositories-InitialSet.txt.zstd[M2]Monitoring/MonitoringExtension.tar.zstd[M2]Monitoring/Monitoring tags.xlsx.zstd[M2]Monitoring/RefactoringMonitor.tar.zstd[M3]StackOverflow/StackOverflow tags.xlsx.zstd[M3]StackOverflow/StackOverflow.tar.zstd
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.018 | 0.076 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.003 |
| Bibliometrics | 0.004 | 0.004 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.003 | 0.004 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.016 | 0.013 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".