Add type stubs

[zzstoatzz]

adds a python/libipld to house the .pyi file and the __all__

following the _{module_name} pattern that I use (and got from pydantic-core), note that maturin develop knows about the module-name set in pyproject.toml

let me know if there are any python type translations I messed up!

before

after

[zzstoatzz]

gentle ping for whenever you have a moment to look @MarshalX - I know there's probably more pressing things so I'll check back later 🙂

[MarshalX]

That's an amazing job, thank you!!

I saw it already but did not comment immediately because this PR also integrates UV. I am still thinking about it but if we want to speed up it let's just include .pyi file :)

[zzstoatzz]

i don't think anything (else, after removing uv.lock) here is uv specific, for example dep groups are a python thing but I can remove it if you want (it shouldn't affect pip setups or anything about your distribution metadata like an extra would)

EDIT: doh the lockfile! duh, let me remove that

[codspeed-hq]

CodSpeed Performance Report

Merging #56 will degrade performances by 19.95%

_{Comparing zzstoatzz:main (a2dba1a) with main (f375d8f)}

Summary

❌ 9 regressions
✅ 183 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
❌	test_dag_cbor_decode_fixtures[dagpb_11unnamedlinks+data]	60.5 µs	68.3 µs	-11.39%
❌	test_dag_cbor_decode_fixtures[float--8.940696716308594e-8]	14.7 µs	16.4 µs	-10.79%
❌	test_dag_cbor_decode_fixtures[int--11959030306112471732]	15.5 µs	17.3 µs	-10.44%
❌	test_dag_cbor_decode_fixtures[int--2784428724]	15.5 µs	17.3 µs	-10.26%
❌	test_dag_cbor_decode_real_data[citm_catalog.json]	24.1 ms	27.4 ms	-12%
❌	test_dag_cbor_decode_real_data[github.json]	1.1 ms	1.3 ms	-14.61%
❌	test_dag_cbor_decode_real_data[twitter.json]	11.3 ms	13.3 ms	-14.57%
❌	test_dag_cbor_decode_torture_cids	71.6 ms	89.4 ms	-19.95%
❌	test_decode_car	694.6 ms	836.9 ms	-17.01%

[zzstoatzz]

Summary

⚡ 6 improvements ❌ 4 regressions ✅ 182 untouched benchmarks

not sure if related to my changes (i suspect not) or if these just have high variance

[MarshalX]

@zzstoatzz that's just a noise from small bechmarks. ignore it :)

[MarshalX]

@zzstoatzz let's give UV a shot! because i do not like this deps list duplication in requirements.txt and in pyproject.toml

[MarshalX]

thanks a lot!! and huge sorry for the so long delay; tbh it was never what to add except stubs to the new release; but now good bug reports came and has been fixed