This is the replication package for the paper entitled "Combining Language and App UI Analysis for the Automated Assessment of Bug Reproduction Steps".

This package contains the following files/folders:

1. a_dataset: a folder containing the required dataset to replicate the results of the paper
   • bug_reports: this folder contains the bug reports used in the study. The bug reports are divided into two sets: development and test sets. The development set contains 54 bug reports, while the test set contains 21 bug reports.
   • execution_model_test_set: this folder contains the application execution models for the 21 bug reports in the test set.
   • execution_model_development_set: this folder contains the application execution models for the 10 bug reports in the development set.
   • QAs: this folder contains quality annotation data
2. b_s2r_sentence_identification: a folder containing the necessary scripts to run the S2R sentence identification phase of AstroBR. This folder has the following sub-folders:
   • generate_prompts: this folder contains the necessary scripts to generate prompts for the S2R sentence identification task.
   • generate_responses: this folder contains the necessary scripts to generate GPT-4 responses for the S2R sentence identification task.
   • result_generation: this folder contains the necessary scripts to generate the results of the S2R sentence identification task. We already provided the necessary result files.
3. c_individual_s2r_extraction: a folder containing the necessary scripts to run the individual S2R extraction phase of AstroBR. This folder has the following sub-folders:
4. d_quality_assessment: a folder containing the necessary scripts to run the quality assessment phase of AstroBR.
   • mapping_gui_response_quality_labels.py: this file will generate annotations for individual S2Rs.
5. e_prompt_templates: a folder containing the prompt templates used in the S2R sentence identification, individual S2R identification, and individual S2R mapping tasks.