Use of generics in requests typing
还没有人认领这个 Issue。
- 主要语言
- Python
- 星标
- 5.1k
- 派生
- 2.1k
- 平均合并
- 1 天 19 小时
- 30 天内合并 PR
- 82
描述
Currently the requests typing makes heavy use of unions. For example the parameter type is:
https://github.com/python/typeshed/blob/f99e8b0615182bb0f17d9d126019d8d72c207ab3/stubs/requests/requests/sessions.pyi#L97-L103
The issue with this approach is that the type system can't keep track of which type has been chosen for a given field. For example, running pyright on this code:
from requests import Session
session = Session()
session.params = {}
reveal_type(session.params)
Shows us that the type system immediately "forgets" that our parameters are definitely in a dictionary:
test.py:5:13 - information: Type of "session.params" is "dict[str | bytes | int | float, str | bytes | int | float | Iterable[str | bytes | int | float] | None]"
0 errors, 0 warnings, 1 information
If we could make Session (as well as other classes such as Request) generic over the parameter type (as well as other fields), then we could keep track of this kind of state. For example, it might look like:
session: Session[dict] = Session()
贡献指南
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
调研方向
先从链接的 requests/requests/sessions.pyi 中第 97-103 行附近的部分开始,然后查看 Session 和 Request 的类型定义方式。使用提供的 pyright 示例来评估泛型参数类型在赋值后是否仍保持具体;当提议的泛型设计已针对受影响的类一致地定义,并保留所选字段类型时,即表示完成。
由索引模型根据 Issue 内容生成。
评估
- 技术栈
- python
- 领域
- tooling
- Issue 类型
- 功能
- 难度
- 5/5
- 预计耗时
- 一周以上
- 活跃度
- 停滞
- 描述清晰度
- 基本清楚
- 新手友好度
- 35/100