Max Klein

Here's a simplified parser implementation based on `ast.parse` that I hacked together: ```python import ast from numpy import s_ def astParseSlice(sliceStr): for node in ast.walk(ast.parse(f's_[{sliceStr}]')): if isinstance(node, ast.ExtSlice): return (x...

Nice, that makes it quite clear how to accomplish this. I'll take a stab at a PR

Sorry, I got a little distracted by some debugging stuff I promised to add to the "how to develop a jupyter extension" jupytercon tutorial. I'll probably make the PR for...

Okay, so I ran into problems with the implementation, and after doing some digging it turns out that the `ast.literal_eval` function changed a bit in 3.8, and the way that...

turns out to have been much simpler than I initially thought to implement a `literal_eval_index` function that works in both cpy37 and cpy38. There's a PR up now with a...

good news: it turns out the code that I already wrote essentially gets support for the new grammar in cpy39 "for free". ast in cpy39 replaced all uses of `ast.Index`...

@mverleg As I say in the README, all that's currently here is an `ndarray` message type. You can use this message type in your own `.proto` files by including `npbuf/type/ndarray.proto`....

@krassowski tree-finder currently supports all of the things in your list in one form or another. There's probably some conceptual gaps, and there's no nice api for a lot of...

@doctorpangloss - good point, I'll add some basic error handling - AFAIK ` importlib.metadata.files` should work in any case where you have a correctly structured/installed Python package. But I've only...

@doctorpangloss > i think you should adopt my approach. Well, I would say that most people think that most of the time :) But still, I liked your suggestions and...