configs.py

import os


class Configs:
    def __init__(self, project_name: str, llm_name: str, quality_level: str = 'L1', prepare_workspace: bool=True):
        self.project_name = project_name
        self.root_dir = os.path.dirname(os.path.realpath(__file__))
        self.quality_level = quality_level
        assert quality_level in ('L1', 'L2', 'L3')

        if llm_name == 'deepseek-v3.1':
            self.openai_key = 'DeeSeekAPIKey'  # deepseek
            self.openai_base_url = 'https://api.deepseek.com'
        else:
            self.openai_key = 'xxxx'  # for GPT
            self.openai_base_url = 'xxxx'

        # repo_name allows mini/variant projects to reuse an existing project's source repo
        self.repo_name = {
            'spark_mini': 'spark',
        }.get(project_name, project_name)

        self.common_project_dir_no_test_file = f'{self.root_dir}/data/repos/repos_removing_test'
        self.common_project_path_no_test_file = f'{self.common_project_dir_no_test_file}/{self.repo_name}'

        self.project_dir = f'{self.root_dir}/data/repos/repos_with_test'
        self.project_with_test_workspace = None

        if prepare_workspace:
            self.prepare_workspace()

        # model configs
        assert llm_name in ['gpt-o4-mini', 'deepseek-v3.1']
        self.llm_name = llm_name
        
        # dataset relevant paths
        if quality_level != 'L1':
            self.coverage_human_labeled_dir = f'{self.root_dir}/data/collected_coverages_{quality_level}'
            q = f'{quality_level}/'
        else:
            self.coverage_human_labeled_dir = f'{self.root_dir}/data/collected_coverages'
            q = ''

        self.test_desc_dataset_path = f'{self.root_dir}/data/test_desc_dataset/{project_name}.json'
        self.fact_set_dir = f'{self.root_dir}/data/fact_set/{project_name}'
        self.dataset_with_generalization_path = f'{self.root_dir}/data/dataset/{project_name}.json'

        # stage-1 relevent paths
        self.generated_exam_save_path = f'{self.root_dir}/data/examination/{q}{llm_name}/{project_name}_exam.json'
        self.generated_exam_answer_save_path = f'{self.root_dir}/data/examination/{q}{llm_name}/{project_name}_answer.json'

        # stage-2 relevant paths
        self.scenario_template_save_path = f'{self.root_dir}/data/generalization/{q}{llm_name}/template_{project_name}.json'
        self.scenario_instances_save_path = f'{self.root_dir}/data/generalization/{q}{llm_name}/instances_{project_name}.json'

        # stage-3 relevant paths
        self.generated_test_case_save_path = f'{self.root_dir}/data/generated_test_cases/{q}{llm_name}/{project_name}.json'
        self.generation_message_log_save_path = f'{self.root_dir}/data/generation_message_log/{q}{llm_name}/{project_name}.json'
        self.test_case_run_log_dir = f'{self.root_dir}/data/test_case_run_log/{q}{llm_name}/{project_name}'
        self.test_case_log_and_coverage_save_path = f'{self.root_dir}/data/generated_test_cases_log_coverage/{q}{llm_name}/{self.project_name}.json'

        self.knowledge_base_dir = f'{self.root_dir}/data/knowledge_base/{project_name}'
        self.knowledge_base_method_path = f'{self.root_dir}/data/knowledge_base/{project_name}/method_def.json'
        self.knowledge_base_method_index_path = f'{self.root_dir}/data/knowledge_base/{project_name}/method_def_index.pt'

        self.knowledge_base_field_path = f'{self.root_dir}/data/knowledge_base/{project_name}/field_def.json'
        self.knowledge_base_field_index_path = f'{self.root_dir}/data/knowledge_base/{project_name}/field_def_index.pt'

        # configurations about program element paths
        self.skip_class_files = True
        
        self.collected_paths_dir = f'{self.root_dir}/data/collected_paths/{project_name}'
        self.collected_paths_focal2method_path = f'{self.collected_paths_dir}/focal_to_method.json'
        self.collected_paths_focal2field_path = f'{self.collected_paths_dir}/focal_to_field.json'

        self.collected_path_dataset_path = f'{self.collected_paths_dir}/dataset.json'
        
        # ablation study
        self.ablate_scenario_template_save_path = f'{self.root_dir}/data/ablate/{llm_name}/template/template_{project_name}.json'
        self.ablate_scenario_instances_save_path = f'{self.root_dir}/data/ablate/{llm_name}/instances/instances_{project_name}.json'
        self.ablate_generated_tests_save_path = f'{self.root_dir}/data/ablate/{llm_name}/tests/generated_tests_{project_name}.json'

        # evaluation on buggy focal methods
        self.buggy_fm_data_raw = f'{self.root_dir}/data/evaluation_on_buggy_fm/dataset_buggy_focal_method/{project_name}_partial_passed.json'
        self.buggy_fm_data_organized = f'{self.root_dir}/data/evaluation_on_buggy_fm/dataset_buggy_focal_method/{project_name}_organized.json'

        self.buggy_generated_exam_save_path = f'{self.root_dir}/data/evaluation_on_buggy_fm/examination/{llm_name}/{project_name}_exam.json'
        self.buggy_generated_exam_answer_save_path = f'{self.root_dir}/data/evaluation_on_buggy_fm/examination/{llm_name}/{project_name}_answer.json'

        self.buggy_scenario_template_save_path = f'{self.root_dir}/data/evaluation_on_buggy_fm/generalization/{llm_name}/template_{project_name}.json'
        self.buggy_scenario_instances_save_path = f'{self.root_dir}/data/evaluation_on_buggy_fm/generalization/{llm_name}/instances_{project_name}.json'
        self.buggy_adjusted_scenario_instances_save_path = f'{self.root_dir}/data/evaluation_on_buggy_fm/generalization/{llm_name}/instances_{project_name}_adjusted.json'

        self.buggy_generated_test_case_save_path = f'{self.root_dir}/data/evaluation_on_buggy_fm/generated_test_cases/{llm_name}/{project_name}.json'

        # project url used for system prompt
        self.project_url = {
            "itext-java": 'https://github.com/itext/itext-java', 
            "hutool": 'https://github.com/chinabugotech/hutool', 
            "yavi": 'https://github.com/making/yavi', 
            "lambda": 'https://github.com/palatable/lambda', 
            "truth": 'https://github.com/google/truth', 
            "cron-utils": 'https://github.com/jmrozanec/cron-utils', 
            "imglib": 'https://github.com/nackily/imglib', 
            "ofdrw": 'https://github.com/ofdrw/ofdrw', 
            "RocketMQC": 'https://github.com/ProgrammerAnthony/RocketMQC', 
            "blade": 'https://github.com/lets-blade/blade', 
            "spark": 'https://github.com/perwendel/spark',
            "spark_mini": 'https://github.com/perwendel/spark',
            "awesome-algorithm": 'https://github.com/codeartx/awesome-algorithm',
            "jInstagram": 'https://github.com/sachin-handiekar/jInstagram'
        }[project_name]

    def prepare_workspace(self):
        # prepare workspace w/o test files for the project
        idx = 0
        while True:
            self.project_dir_no_test_file = f'{self.common_project_dir_no_test_file}/workspace_{idx}'
            self.project_path_no_test_file = f'{self.project_dir_no_test_file}/{self.project_name}'
            if not os.path.exists(self.project_path_no_test_file):
                os.makedirs(self.project_dir_no_test_file, exist_ok=True)
                os.system(f'cp -r {self.common_project_path_no_test_file} {self.project_path_no_test_file}')
                break
            idx += 1

        # prepare workspace with test files for the project (evaluation purpose)
        idx = 0
        while True:
            project_dir_with_test_file = f'{self.project_dir}/workspace_{idx}'
            project_path_with_test_file = f'{project_dir_with_test_file}/{self.project_name}'
            if not os.path.exists(project_path_with_test_file):
                os.makedirs(project_dir_with_test_file, exist_ok=True)
                os.system(f'cp -r {self.project_dir}/{self.repo_name} {project_path_with_test_file}')
                self.project_with_test_workspace = project_dir_with_test_file
                break
            idx += 1