Python如何通过读写txt文件实现多脚本按状态分支执行

该功能完全可以实现，也是文字互动类游戏最常用的跨次启动状态持久化方案，你的核心思路正确，只是现有代码的文件操作和流程逻辑存在几个容易触发bug的细节问题，修正后即可正常运行。

• 文件操作逻辑存在严重隐患：全局以r+模式打开文件句柄，用f.seek(2)偏移固定字节读取内容的方式非常脆弱，只要文件内容长度不符合预期就会读错值；另外全局打开的文件默认不会实时把写入内容刷到磁盘，你写完状态立刻杀进程的话，大概率内容还留在内存缓冲区，下次启动根本读不到新写入的章节标识。
• 外部脚本加载方式不规范：用exec(open("chapter1.py").read())执行外部脚本，会把外部脚本的所有变量、函数直接注入当前主程序的命名空间，后续章节多了很容易出现变量名冲突，排查问题非常麻烦。
• 初始状态逻辑有漏洞：第一次启动检测到文件为空写入"0"后，没有触发初始章节流程，程序会直接退出。
• 没有做脏数据兼容：如果存档文件被手动修改、写入了非数字内容，现有逻辑会直接失效，没有兜底的初始状态判断。

把文件读写逻辑拆成独立函数，每次读写单独打开文件、用完即关，写入时强制刷盘保证内容不丢失；章节加载改用模块导入的方式，避免命名空间污染；去掉不稳定的seek偏移，直接读取整个文件内容做解析。

from time import sleep
import os

# 存档文件路径，和脚本放同目录即可，不用写死绝对路径
SAVE_PATH = "chapter.txt"

def read_state():
    """读取存档状态，文件不存在或内容非法时默认返回初始状态0"""
    if not os.path.exists(SAVE_PATH):
        return 0
    with open(SAVE_PATH, "r", encoding="utf-8") as f:
        content = f.read().strip()
    # 只识别第一个有效数字，兼容手动修改文件带来的脏数据
    if content and content[0].isdigit():
        return int(content[0])
    return 0

def write_state(state_num):
    """写入章节状态，覆盖式写入保证文件内仅存当前状态值"""
    with open(SAVE_PATH, "w", encoding="utf-8") as f:
        f.write(str(state_num))
        # 强制刷写系统缓冲区，确保杀进程前内容已经写入磁盘
        f.flush()
        os.fsync(f.fileno())

def intro():
    print(". . .\n")
    sleep(3)
    print("What are you doing?\n")
    sleep(3)
    print("There is nothing here, go away.\n")
    sleep(5)
    print("You're not going to leave, huh?\n")
    sleep(5)
    print("What if I told you that this is not a program?\n")
    sleep(8)
    print("Nothing, huh?\n")
    sleep(4)
    print("Oh, that's right.  You can't type.  Here let me fix that.\n")
    sleep(9)
    print("Almost there. . .\n")
    sleep(4)
    print("and. . .\n")
    sleep(3)
    print("Done\n")
    sleep(3)
    ans = input("Type \"yes\" if you are able to type now: ")
    if ans == "yes":
        sleep(2)
        print("\nOk, great!\n")
        sleep(4)
        if_yes()
    elif ans == "no":
        sleep(3)
        print("\n. . .\n")
        sleep(3)
        if_no()

def if_yes():
    reason = input("Now that you are able to type, why are you still here? (Curiosity or irritation): ")
    if reason == "curiosity":
        sleep(1.5)
        print("\nWell, you don't have to be.  Because there is nothing for you here.\n")
        sleep(2)
        print("So just LEAVE.\n")
        for i in range(3):
            sleep(10)
            print(". . .\n")
        sleep(5)
        print("Ok, fine.\n")
        sleep(2)
        print("You want something to happen?\n")
        sleep(1.5)
        print("I warned you.\n")
        sleep(3)
        print("~~~Not_Program has been forcefully terminated~~~")
        sleep(2)
        # 必须先写存档，再执行杀进程操作
        write_state(1)
        os.system("taskkill /f /im cmd.exe")
    elif reason == "irritation":
        print("\nWell that's just mean.\n")
        sleep(2)
        print("Do you enjoy hurting me?\n")
        sleep(2)
        print("I'm just a prog-I mean person, just like you.\n")
        sleep(2)
        print(". . .\n")
        sleep(2)
        print("What is this?\n")
        sleep(1)
        print("Am I just your entertainment?\n")
        sleep(2.5)
        print("Well, I won't indulge you.\n")
        sleep(3)
        print("~~~Not_Program has been forcefully terminated~~~")
        sleep(2)
        write_state(1)
        os.system("taskkill /f /im cmd.exe")

def if_no():
    print("Really?\n")
    sleep(4)
    print("Fine, you want to be like that?\n")
    sleep(3)
    print("Well, you asked for it.\n")
    sleep(5)
    print("~~~Not_Program has been forcefully terminated~~~")
    sleep(2)
    write_state(1)
    os.system("taskkill /f /im cmd.exe")

def load_chapter(chap_num):
    """根据状态值加载对应章节，后续新增章节直接加分支即可"""
    if chap_num == 0:
        intro()
    elif chap_num == 1:
        # 推荐方式：章节脚本内定义统一入口函数run，导入后调用，避免变量冲突
        import chapter1
        chapter1.run()
        # 临时过渡也可以用exec，注意指定编码避免乱码
        # exec(open("chapter1.py", "r", encoding="utf-8").read())

if __name__ == "__main__":
    current_state = read_state()
    load_chapter(current_state)

• 写入顺序不能乱：所有触发强制退出的分支，必须先调用write_state写入新的章节状态，再执行杀进程的操作，否则会出现存档丢失的问题。
• 存档文件始终用覆盖模式写入，不要追加，避免文件内堆积多个数字导致读取错误。
• 后续新增章节只需要按顺序分配数字状态（2、3对应第二章、第三章），在load_chapter函数里新增对应判断分支即可，逻辑和第一章完全一致。
• 如果需要做跨平台兼容，不要硬编码Windows专属的taskkill命令，当前Windows环境下使用无影响。

内容的提问来源于stack exchange，提问作者Tyler Slade