4.1 文本文件读写

推荐写法：with open(…)

with open("path.txt", "r", encoding="utf-8") as f:
    content = f.read()           # 全部读成字符串
    lines = f.readlines()        # 按行读成 list

with open("out.txt", "w", encoding="utf-8") as f:
    f.write("hello")
    f.writelines(["a\n", "b\n"])

关键点

with 块结束时会自动关文件（即使中途异常）
encoding="utf-8" 必须显式写，否则不同平台行为不一致
模式："r" 读、"w" 写（覆盖）、"a" 追加

为什么推理代码也用

读 system prompt、读模型配置文件、读 vocab、读 few-shot 例子。

示范

# 写
with open("prompt.txt", "w", encoding="utf-8") as f:
    f.write("You are a helpful assistant.\n")

# 读
with open("prompt.txt", "r", encoding="utf-8") as f:
    print(repr(f.read()))

✍️ 练习

写代码：

用 with open(...) 把字符串 "Qwen2.5-7B-Instruct\nvllm-0.6\n" 写到文件 model_card.txt
再用 with open(...) 把它读出来，赋值给 content，然后 print(content

注意编码用 utf-8，写模式用 "w"，读模式用 "r"。

💡 思路提示

点开看提示

with open(path, mode, encoding="utf-8") as f: 是固定模式
读全部内容用 f.read()
print(repr(content)) 能看到 \n 是否真的存在

✅ 参考解法

写不出来再打开

path = "model_card.txt"

with open(path, "w", encoding="utf-8") as f:
    f.write("Qwen2.5-7B-Instruct\nvllm-0.6\n")

with open(path, "r", encoding="utf-8") as f:
    content = f.read()

print(content)

🔍 进阶思考

f.read() 把整个文件读成一个大字符串；超大文件用 for line in f: 一行行读，内存友好。

with open("huge.log", "r", encoding="utf-8") as f:
    for line in f:
        process(line)