testcase-data/eval/right_code_experiment.py

import json

exe_results = "/home/i-luoxianzhen/data/TestCase-Gen/eval/ALLmode_results/tcb-lcb-rank5-all.json"

results = json.load(open(exe_results, "r", encoding="utf-8"))

status_count = {}
ALL_status = {
    0: {"AC": 0, "CE": 0, "TLE": 0, "MLE": 0, "RE": 0, "WA": 0, "EXE": 0},
    1: {"AC": 0, "CE": 0, "TLE": 0, "MLE": 0, "RE": 0, "WA": 0, "EXE": 0},
    2: {"AC": 0, "CE": 0, "TLE": 0, "MLE": 0, "RE": 0, "WA": 0, "EXE": 0},
    3: {"AC": 0, "CE": 0, "TLE": 0, "MLE": 0, "RE": 0, "WA": 0, "EXE": 0},
    4: {"AC": 0, "CE": 0, "TLE": 0, "MLE": 0, "RE": 0, "WA": 0, "EXE": 0},
    5: {"AC": 0, "CE": 0, "TLE": 0, "MLE": 0, "RE": 0, "WA": 0, "EXE": 0},
    6: {"AC": 0, "CE": 0, "TLE": 0, "MLE": 0, "RE": 0, "WA": 0, "EXE": 0},
    7: {"AC": 0, "CE": 0, "TLE": 0, "MLE": 0, "RE": 0, "WA": 0, "EXE": 0},
    8: {"AC": 0, "CE": 0, "TLE": 0, "MLE": 0, "RE": 0, "WA": 0, "EXE": 0},
}
test_dir = "/home/i-luoxianzhen/data/TestCase-Gen/save_tests_scaling/lcb/tests-{}.jsonl"
for k, v in results.items():
    status_count[k] = {}
    saved_nums = 0
    for code_id, item in enumerate(v['codes']):
        status_count[k][code_id] = {"AC": 0, "CE": 0, "TLE": 0, "MLE": 0, "RE": 0, "WA": 0, "EXE": 0}
        for status in item['status']:
            status_count[k][code_id][status] += 1
            ALL_status[code_id][status] += 1
                

columns = ["Index"] + list(next(iter(ALL_status.values())).keys())

# 生成 Markdown 表格
header = "| " + " | ".join(columns) + " |"
separator = "| " + " | ".join(["---"] * len(columns)) + " |"
rows = []
for idx, values in ALL_status.items():
    row = "| " + str(idx) + " | " + " | ".join(str(values[col]) for col in columns[1:]) + " |"
    rows.append(row)

# 拼接最终结果
markdown_table = "\n".join([header, separator] + rows)

print(markdown_table)