generate_test_data 源代码

"""
生成测试数据脚本
复制 Ren'py.xlsx 并填充 1000 行随机测试数据
"""
import random
import shutil
from pathlib import Path
from openpyxl import load_workbook
from openpyxl.worksheet.datavalidation import DataValidation



def generate_test_data():
    """生成测试数据"""
    # 源文件和目标文件
    source_file = Path("input/Ren'py.xlsx")
    target_file = Path("input/Ren'py_2.xlsx")

    # 复制文件
    print(f"复制文件: {source_file} -> {target_file}")
    shutil.copy(source_file, target_file)

    # 打开目标文件
    wb = load_workbook(target_file)
    ws = wb.active

    # 获取参数表工作表（用于获取数据验证列表）
    param_ws = wb["参数表"] if "参数表" in wb.sheetnames else None

    # 读取参数表的列名和数据
    param_data = {}
    if param_ws:
        for col_idx in range(1, param_ws.max_column + 1):
            header = param_ws.cell(row=1, column=col_idx).value
            if header:
                values = []
                for row_idx in range(2, param_ws.max_row + 1):
                    val = param_ws.cell(row=row_idx, column=col_idx).value
                    if val:
                        values.append(str(val))
                param_data[header] = values
        print(f"从参数表读取到 {len(param_data)} 个参数类型")

    # 获取主工作表的表头
    headers = []
    for col_idx in range(1, ws.max_column + 1):
        header = ws.cell(row=1, column=col_idx).value
        if header:
            headers.append((col_idx, header))

    print(f"主工作表有 {len(headers)} 列")

    # 生成 1000 行随机数据
    start_row = ws.max_row + 1
    print(f"从第 {start_row} 行开始生成 1000 行数据...")

    for row_num in range(start_row, start_row + 1000):
        for col_idx, header in headers:
            # 根据列名生成合适的随机数据
            if header == "Index":
                # Index 列递增
                value = row_num - 1
            elif header == "Ignore":
                # Ignore 列保持为空
                value = ""
            elif header in param_data and param_data[header]:
                # 从参数表中随机选择
                value = random.choice(param_data[header])
            elif header == "Speaker":
                value = random.choice(["角色A", "角色B", "角色C", "旁白", ""])
            elif header == "Text":
                value = f"这是第{row_num}行的测试文本内容。"
            elif header in ["Wait", "Visible", "CharAnimWait"]:
                value = random.choice(["", "true", "false"])
            elif header in ["Volume", "Scale", "Time"]:
                value = random.choice(["", "0.5", "1.0", "1.5"])
            elif header in ["Position"]:
                value = random.choice(["", "0.3", "0.5", "0.7"])
            else:
                # 其他列随机留空或填充
                value = "" if random.random() > 0.3 else f"test_{row_num}"

            ws.cell(row=row_num, column=col_idx, value=value)

        if row_num % 100 == 0:
            print(f"  已生成 {row_num - start_row + 1} 行...")

    # 保存文件
    print(f"保存文件: {target_file}")
    wb.save(target_file)
    wb.close()

    print(f"完成！生成了 1000 行测试数据")
    print(f"文件总行数: {start_row + 1000 - 1} 行（含表头）")


if __name__ == "__main__":
    generate_test_data()