Auto generate pyright test config

Take the information which stubs to exclude from the METADATA.toml files. On one hand this is necessary for #10722, i.e. to correctly handle requires_python, on the other hand this allows us to move a "stub incomplete" marker into the METADATA.toml file, and use it for other purposes than just the pyright tests. E.g. instead of relying solely on pyright to prevent regressions of "complete" stubs (which doesn't work great, since pyright doesn't understand Incomplete), we could have a separate script in CI.

首先检查 METADATA.toml 文件和当前的 pyright 测试配置，然后查看 issue #10722 中的上下文。当相关的存根排除信息和不完整存根信息已在 METADATA.toml 中体现，并且能够支持 pyright 测试和其他 CI 检查时，即视为完成。