enderpy VS pylyzer

Note on LSP servers Usually when using VSCode, you just select a language and install the corresponding language server extension from Marketplace, e.g. Pyright of Ruff for Python. Under the hood, most of these VSCode language server extensions utilize vscode-languageclient api. The API allows to launch LSP server in several ways, e.g. as a node module running in runtime provided by VSCode itself, or as a child process via runnable command. You can take a look at the Pylyzer Python LSP extension to see an example of usage of the API. Note that in order to use it, you need a runtime that has access to your files. There is a possibility to add VSCode server to your Monaco project and use it to launch language servers, however it adds additional complexity and dependency. In this guide I will avoid it. There are other ways to run Language Server, e.g. one can create a new language server or a wrapper for existing one with pygls, to run it as Python process providing websocket server. Here is a great guide with introduction to language servers and monaco language client. Other similar option but for Rust is tower-lsp.

3.pylyzer

There's pylyzer[0], but it's in the early stages.

[0] https://github.com/mtshiba/pylyzer

I’ve just found out pylyzer https://github.com/mtshiba/pylyzer

Not sure how good it is.

That explains why I didn't see textDocument/semanticTokens/full or anything like it in pyright's code. Do you happen to know of any Python LSP that has semantic tokens? pylsp, pyright, and jedi-language-server all don't have it. The only one I could find that might support it is https://github.com/mtshiba/pylyzer but I haven't tried that LSP yet

You might want to check Pylyzer then (https://github.com/mtshiba/pylyzer).

I'm not involved at all. It is still very very early in development. But as it is in the same vein, I thought I'd mentioned it here.

I don't need a lot. Something like "go to definition" and "show docstring" would go a long way, "signature help" would be nice... I'd take a nice completion, but that would just be a cherry on top. Any low-effort ways to achieve this? I've found https://github.com/mtshiba/pylyzer which looked really nice (and I could just download the right binary), but it has another programming language as a dependency, which is kind of intimidating...

Perhaps pylyzer can deliver. https://github.com/mtshiba/pylyzer