Announcing the Beta release of ty
astral.sh854 points by gavide 4 months ago
854 points by gavide 4 months ago
Hopefully it gets added to this comparison:
https://htmlpreview.github.io/?https://github.com/python/typ...
If that table is anything to go by, Pyright is not to be underestimated.
I have briefly tried ty (LSP) in Emacs and it seems to work well so far. The only questionable thing I've encountered is that when the signature of a method is shown, the type annotations of some parameters seem to be presented in a particularly verbose form compared to what I'm used to - maybe they're technically correct but it can be bit much to look at.
Anyway, odds are pretty good that ty is what I will end up using long-term, so thanks and congrats on releasing the first beta!
Note: while spec conformance is important, I don't recommend using it as the basis for choosing a type checker. It is not representative of the things that most users actually care about (and is not meant to be).
(I was on the Python Typing Council and helped put together the spec, the conformance test suite, etc)
Can you add some examples of the things users care about that aren't well covered by this? I empathize with everyone who wants a feature comparison chart so they can be confident switching without unknowingly losing important safety checks.
The conformance test suite is currently mostly focused on “what does an explicit type annotation mean”
A shared spec for this is important because if you write a Python library, you don’t want to have to write a different set of types for each Python type checker
Here are some things the spec has nothing to say about:
Inference
You don’t want to annotate every expression in your program. Type checkers have a lot of leeway here and this makes a huge difference to what it feels like to use a type checker.
For instance, mypy will complain about the following, but pyright will not (because it infers the types of unannotated collections as having Any):
x = []
x.append(1)
x[0] + "oops"
Diagnostics
The spec has very little to say about what a type checker should do with all the information it has. Should it complain about unreachable code? Should it complain if you did `if foo` instead of `if foo()` because it’s always true? The line between type checker and linter is murky. Decisions here have nothing to do with “what does this annotation mean”, so are mostly out of scope from the spec.
Configuration
This makes a huge difference when adapting huge codebases to different levels of type checking. Also the defaults really matter, which can be tricky when Python type checkers serve so many different audiences.
Other things the spec doesn’t say anything about: error messages quality, editor integration, speed, long tail of UX issues, implementation of new type system features, plugins, type system extensions or special casing
And then of course there are things we would like to spec but haven’t yet!
> but pyright will not (because it infers the types of unannotated collections as having Any)
This is incorrect. pyright will infer the type of x as list[Unknown].
Unknown has the exact same type system semantics as Any.
Unknown is a pyright specific term for inferred Any that is used as the basis for enabling additional diagnostics prohibiting gradual typing.
Notably, this is quite different from TypeScript’s unknown, which is type safe.
In case you’re not well versed in Python typecheckers, in the mypy vs Pyright example, Pyright can be configured to complain about not annotating the collection (and so both typecheckers will yell at the code as written).
TypeScript takes the same approach in this scenario, and I assume this helps both be fast.
They were "on the Python Typing Council and helped put together the spec, the conformance test suite, etc" so I assume they are an expert on Python typecheckers
I didn’t write it for parent lol. I guess I should be more careful with “you”.
TypeScript will use flow typing to determine the type as number[] in this code:
const x = []
x.push(1)
type t = typeof x // number[]I think the idea is not that there are features that aren’t listed, but rather that if a typechecker supports 10 features people care about and is missing 10 that people don’t really use, it will look a lot worse on a list like this than a typechecker with 100% compliance, when in practice it may not really be worse at all.
Edit: Based on this other comment, the point was also about things not covered by the spec. “The spec mostly concerns itself with the semantics of annotations, not diagnostics or inference.” https://news.ycombinator.com/item?id=46296360
The chart does not describe speed (either in general or in any particular case). Speed/performance/latency is a thing users care about that is not included in the feature list.
Yea that one is fine and well covered in the blog post, and pretty easy to spot in light testing. I'm much more worried about the ones that are harder to spot until you have a false negative that turns into a real bug which would be caught by 1 tool and not another.
I can't recall a single time that type-checker speed was the limiting factor for me.
I can't say I've been bottlenecked on it, but I've certainly been bothered by it.
We'll be adding ourselves to that table soon. We'll have some work to catch up with pyright on conformance, but that's what the time between now and stable release is for.
pyright is very good, but there is also https://docs.basedpyright.com/latest/ which improves on it further.
That said I'm very happy user of uv, so once Ty becomes ready enough will be happy to migrate.
Basedpyright plus any AI generated python is a hellscape unless you use hooks and have a lot of patience.
Not sure where the AI generated python is coming from?
AI generated anything is a hellscape.
Pyright is really really good. Anyone that doubts that 10x engineers exist, just go and look at Eric Traut. He's pretty much written it single handedly. Absolute machine.
Mypy is trash. Nice to have a table to point to to prove it.
Oh my, I just looked him up. He is the developer of Virtual Game Station - a PS1 emulator that I used in the past to play PS Isos on my Windows ME PC! What a legend.
Mypy is not trash at all. It is still the most precise type checker, supports a lot of complicated cases and has plugins that make a lot of things more type safe. I believe Mypy is severely underrated and a lot of people believe that it's not as good as Pyright because of the conformance tests.
The problem is that the conformance tests were mostly written by Eric Traut, so there's a natural bias towards specifying what Pyright does well. There's a lot of things Mypy does really well that should probably be implemented in Pyright.
Pyright has been great. But it’s slow. Speed of a LSP does matter for UX. Excited to see how much ty improves on this.
Is it wrong to to say that I don't like pyright on principle because it requires node.js and npm to install and run?
I think it is way to slow too. The one from microsoft (pylance IIRC) is better in my opinion.
Pylance's type checker is Pyright, so in that particular respect they're exactly the same.
Pyright is a type checker, not a LSP per se in my opinion. ty is both.
https://github.com/python/typing/pull/2137
PR is somewhat WIP-ish but I needed some motivation to do OSS work again :)
For those interested: the results page in this PR looks like this:
https://htmlpreview.github.io/?https://github.com/SimonSchic...
Thanks.
I really need better generics support before ty becomes useful. Currently decorators just make all return types unknown. I need something this to work:
_F = TypeVar("_F", bound=Callable[..., Any])
def my_decorator(*args: str) -> Callable[[_F], _F]: ...
I really hope astral can monetize without a highly destructive rugpull, because they are building great tools and solving real problems.
"pyx" is their first commercial offering: https://astral.sh/pyx
I agree though. Hope this is successful and they keep building awesome open-source tools.
We're paying for pyx. Wouldn't have if we didn't enjoy enjoy uv and ruff.
It's definitely a narrow path for them to tread. Feels like the best case is something like Hashicorp, great until the founders don't want to do it anymore.
> Feels like the best case is something like Hashicorp
Wow, that's probably my go-to case of things going south, not "best case scenario". They sold to IBM, a famous graveyard for software, and on the way there changed from FOSS licensing to their own proprietary ones for software the community started to rely on.
You're not wrong, but a) most of the badness happened after the founders checked out and b) it's hard to find examples of developer tool companies doing better.
You however, are. Hashimoto didn't leave until December 2023, Hashicorp announced the license change August 10, 2023. Also way back in September 2021 they started having staffing issues and stopped accepting community contributions, and also made the questionable choice of going public that same year.
You might be on to something with point B, hard to find good examples of developer tool companies that don't eventually turn sour. However, there are countless examples of successful and still very useful developer tools out there, maybe slapping a company on it and sell a "pro" version isn't the way to go?
I actually would argue that Hashimoto "left" earlier. He "stepped down" from the executive team July 2021 and became an individual contributor then. He likely lost interest/power a long time before 2023.
https://www.hashicorp.com/en/blog/mitchell-s-new-role-at-has...
Meh. The end of the company many of us admire was a combination of the founders giving up control to the usual villains and the venture business model failing for developer tools. I don't think the specific departure date matters very much; things started to degrade earlier.
As for "slapping a company on it", I agree, but also I don't think we've developed a viable alternative. Python has been limping along with one toolchain or another for my entire career (multiple decades) and it took Astral's very specific approach to create something better. It's fair to ask why they needed to be venture backed, but they clearly are and the lack of successful alternatives is telling.
Yeah their work thus far has been an incredible public service to the Python community.
Feels like they’re headed in the direction of bun.
In zero revenue or acquisition direction
Thankfully all these LLM labs are heavily invested in python so this seems like the likely route IMO
My issue with them is that they claim their tools replace existing tools, but they don't bother to actually replicate all of the functionality. So if you want to use the full functionality of existing tools, you need to fall back on them instead of using Astral's "replacements". It's like one step forward and one step back. For me personally, speed of the tooling is not as important as what the tooling can check, which is very important for a language like Python that is very easy to get wrong.
If there are specific incompatibilities or rough edges you're running into, we're always interested in hearing about them. We try pretty hard to provide a pip compatibility layer[1], but Python packaging is non-trivial and has a lot of layers and caveats.
Is there any plan for a non-“compatibility layer” way to do anything manual or nontrivial? uv sync and uv run are sort of fine for developing a distribution/package, but they’re not exactly replacements for anything else one might want to do with the pip and venv commands.
As a very basic example I ran into last week, Python tooling, even the nice Astral tooling, seems to be almost completely lacking any good detection of what source changes need to trigger what rebuild steps. Unless I’ve missed something, if I make a change to a source tree that uv sync doesn’t notice, I’m stuck with uv pip install -e ., which is a wee bit disappointing and feels a bit gross. I suppose I could try to put something correct into cache-keys, but this is fundamentally wrong. The list of files in my source tree that need to trigger a refresh is something that my build system determines when it builds. Maybe there should be a way to either plumb that into uv’s cache or to tell uv that at least “uv sync” should run the designated command to (incrementally) rebuild my source tree?
(Not that I can blame uv for failing to magically exfiltrate metadata from the black box that is hatchling plus its plugins.)
> Is there any plan for a non-“compatibility layer” way to do anything manual or nontrivial?
It's really helpful to have examples for this, like the one you provide below (which I'll respond to!). I've been a maintainer and contributor to the PyPA standard tooling for years, and once uv "clicked" for me I didn't find myself having to leave the imperative layer (of uv add/sync/etc) at all.
> As a very basic example I ran into last week, Python tooling, even the nice Astral tooling, seems to be almost completely lacking any good detection of what source changes need to trigger what rebuild steps.
Could you say more about your setup here? By "rebuild steps" I'm inferring you mean an editable install (versus a sdist/bdist build) -- in general `uv sync` should work in that scenario, including for non-trivial things where e.g. an extension build has to be re-run. In other words, if you do `uv sync` instead of `uv pip install -e .`, that should generally work.
However, to take a step back from that: IMO the nicer way to use uv is to not run `uv sync` that much. Instead, you can generally use `uv run ...` to auto-sync and run your development tooling within an environment than includes your editable installation.
By way of example, here's what I would traditionally do:
python -m venv .env
source .env/bin/activate
python -m pip install -e .[dev] # editable install with the 'dev' extra
pytest ...
# re-run install if there are things a normal editable install can't transparently sync, like extension builds
uv run --dev pytest ... # uses pytest from the 'dev' dependency group
My setup is a mixed C/C++/Python project. The C and C++ code builds independently of the Python code (using waf, but I think this barely matters -- the point is that the C/C++ build is triggered by a straightforward command and that it rebuilds correctly based on changed source code). The Python code depends on the C/C++ code via ctypes and cffi (which load a .so file produced by the C/C++ build), and there are no extension modules.
Python builds via [tool.hatch.build.targets.wheel.hooks.custom] in pyproject.toml and a hatch_build.py that invokes waf and force-includes the .so files into useful locations.
Use case 1: Development. I change something (C/C++ source, the waf configuration, etc) and then try to run Python code (via uv sync, uv run, or activating a venv with an editable install). Since there doesn't seem to be a way to have the build feed dependencies out to uv (this seems to be a deficiency in PEP 517/660), I either need to somehow statically generate cache-keys or resort to reinstall-package to get uv commands to notice when something changed. I can force the issue with uv pip install -e ., although apparently I can also force the issue with uv run/sync --reinstall-packages [distro name]. [0] So I guess uv pip is not actually needed here.
It would be very nice if there was an extension to PEP 660 that would allow the editable build to tell the front-end what its computed dependencies are.
Use case 2: Production
IMO uv sync and uv run have no place in production. I do not want my server to resolve dependencies or create environments at all, let alone by magic, when I am running a release of my software built for the purpose.
My code has, long before pyproject.toml or uv was a thing and even before virtual environments existed (!), had a script to build a production artifact. The resulting artifact makes its way to a server, and the code in it gets run. If I want to use dependencies as found by uv, or if I want to use entrypoints (a massive improvement over rolling my own way to actually invoke a Python program!), as far as I can tell I can either manually make and populate a venv using uv venv and uv pip or I can use UV_PROJECT_ENVIRONMENT with uv sync and abuse uv sync to imperatively create a venv.
Maybe some day uv will come up with a better way to produce production artifacts. (And maybe in the distant future, the libc world will come up with a decent way to make C/C++ virtual environments that don't rely on mount namespaces or chroot.)
[0] As far as I can tell, the accepted terminology is that the thing produced by a pyproject.toml is possibly a "project" or a "distribution" and that these are both very much distinct from a "package". I think it's a bit regrettable that uv's option here is spelled like it rebuilds a _package_ when the thing you feed it is not the name of a package and it does not rebuild a particular package. In uv's defense, PEP 517 itself seems rather confused as well.
uv needs to support creation of zipapps, like pdm does (what pex does standalone).
Various tickets asking for it, but they also want to bundle in the python interpreter itself, which is out of scope for a pyproject.toml manager: https://github.com/astral-sh/uv/issues/5802
Their integration with existing tools seems to be generally pretty good.
For example, uv-build is rather lacking in any sort of features (and its documentation barely exists AFAICT, which is a bit disappointing), but uv works just fine with hatchling, using configuration mechanisms that predate uv.
(I spent some time last week porting a project from an old, entirely unsupportable build system to uv + hatchling, and I came out of it every bit as unimpressed by the general state of Python packaging as ever, but I had no real complaints about uv. It would be nice if there was a build system that could go even slightly off the beaten path without writing custom hooks and mostly inferring how they’re supposed to work, though. I’m pretty sure that even the major LLMs only know how to write a Python package configuration because they’ve trained on random blog posts and some GitHub packages that mostly work — they’re certainly not figuring anything out directly from the documentation, nor could they.)
Getting from 95% compatible to 100% compatible may not only take a lot of time, but also result in worsening the performance. Sometimes it's good to drop some off the less frequently used features in order to make the tool better (or allow for making the tool better)
Damn it, this unicorn farting rainbows and craping gold is not yet capable of towing another car. I don't know why they advertise it as a replacement for my current mode of transportation.
Thanks Astral team! We use Pydantic heavily, and it looks like first class support from Ty is slated for the stable release, we'd love to try it.
While we wait... what's everyone's type checking setup? We run both Pyright and Mypy... they catch different errors so we've kept both, but it feels redundant.
https://htmlpreview.github.io/?https://github.com/python/typ... suggests that Pyright is a superset, which hasn't matched our experience.
Though our analysis was ~2 years ago. Anyone with a large Python codebase successfully consolidated to just Pyright?
I appreciate the even tempered question. I’ve been using mypy since its early days, and when pyright was added to vs code I was forced to reckon with their differences. For the most part I found mypy was able to infer more accurately and flexibly. At various times I had to turn pyright off entirely because of false positives. But perhaps someone else would say that I’m leaning on weaknesses of mypy; I think I’m pretty strict but who knows. And like yourself, mine is a rather dated opinion. It used to be that every mypy release was an event, where I’d have a bunch of new errors to fix, but that lessened over the years.
I suspect pyright has caught up a lot but I turned it off again rather recently.
For what it’s worth I did give up on cursor mostly because basedpyright was very counterproductive for me.
I will say that I’ve seen a lot more vehement trash talking about mypy and gushing about pyright than vice versa for quite a few years. It doesn’t quite add up in my mind.
I’ve added ecosystem regression checks to every Python type checker and typeshed via https://github.com/hauntsaninja/mypy_primer. This helped a tonne with preventing unintended or overly burdensome regressions in mypy, so glad to hear upgrades are less of an Event for you
> I will say that I’ve seen a lot more vehement trash talking about mypy and gushing about pyright than vice versa for quite a few years. It doesn’t quite add up in my mind.
agreed! mypy's been good to us over the years.
The biggest problem we're looking to solve now is raw speed, type checking is by far the slowest part of our precommit stack which is what got us interested in Ty.
Mentioned this in another comment, but the spec conformance suite is not representative of the things users care about (nor is it meant to be).
The spec mostly concerns itself with the semantics of annotations, not diagnostics or inference. I don't really recommend using it as the basis for choosing a type checker.
(I was on the Python Typing Council and helped put together the spec, the conformance test suite, etc)
The title of this story should be "Announcing the Beta release of ty". A lot of people have been waiting for the beta specifically.
I've been using Pyrefly and loving it compared to Pyright, but they recently shipped some updates with crash bugs that forced me to pin to a previous version, which is annoying. Unfortunately my first impression of ty isn't great either. Trying to install the ty extension on the current version of Cursor says "Can't install 'astral-sh.ty' extension because it is not compatible with the current version of Cursor (version 2.2.20, VSCode version 1.105.1)."
(pyrefly maintainer here) If you haven't already, please file an issue for that crash on the [Pyrefly repo](https://github.com/facebook/pyrefly) as well :)
If there's anything else accompanying the error, do you mind filing an issue? I've been using the ty extension with Cursor for weeks and am having trouble reproducing right now.
That's the full error. It shows up in a dialog box when I press the install button. I'm on macOS, connected with the Anysphere Remote SSH extension to a Linux machine.
If I choose "install previous version" I am able to install the pre-release version from 12 hours ago without issue. Then on the extension page I get a button labeled "Switch to Release Version" and when I press it I get an error that says "Can't install release version of 'ty' extension because it has no release version." Filed a GitHub issue with these details.
In the meantime, the previous version appears to be working well! I like that it worked without any configuration. The Pyrefly extension needed a config tweak to work.
https://forum.cursor.com/t/newly-published-extensions-appear... suggests that there's some kind of delayed daily update for new VSCode extension versions to become available to Cursor? It seems likely that's what is happening here, since ty-vscode 0.0.2 was only published an hour or two ago.
Oh, huh, and since there's no previous release version it just fails to install completely? That's unfortunate.
Apart from installation problems/crash issues, do you have some feedback about type checking with ty vs. pyrefly? Which is stricter, soundness issues, etc?
Both are rust/open-source/new/fast so it's difficult to understand why I should choose one over the other.
I still don’t understand how a single language can have multiple (what is it now, half a dozen?) different type checkers, all with different behaviour.
Do library authors have to test against every type checker to ensure maximum compatibility? Do application developers need to limit their use of libraries to ones that support their particular choice of type checker?
You’re talking about a duck typed language with optional type annotations. I love python but that’s a combination that should explain a bit why there are so many different implementations.
It doesn't. Either the optional type annotations have precise semantics or they don't.
The annotations have fairly well defined semantics, the behavior of typecheckers in the absence of annotations, where types are ambiguous (a common case being when the type is a generic collection type but the defining position is assignment to an empty collection so that the correct specialization of the generic type is ambiguous) is less defined.
What should a type checker say about this code?
x = []
x.append(1)
x[0] = "new"
x[0] + "oops"
TypeScript widens the type of x to allow `number | string`, there are no type errors below:
const x = []
x.push(1)
type t = typeof x
// ^? type t = number[]
x[0] = "new"
type t2 = typeof x
// ^? type t2 = (number | string)[]
const y = x[0] + "oops"
// ^? const y: string
It depends on the semantics the language specifies. Whether or not the annotations are optional is irrelevant.
Either way, you didn't annotate the code so it's kind of pointless to discuss.
Also fwiw python is typed regardless of the annotations; types are not optional in any sense. Unless you're using BCPL or forth or something like that
> Either way, you didn't annotate the code so it's kind of pointless to discuss.
There are several literals in that code snippet; I could annotate them with their types, and this code would still be exactly as it is. You asked why there are competing type checkers, and the fact that the language is only optionally typed means ambiguity like that example exists, and should be a warning/bug/allowed; choose the type checker that most closely matches the semantics you want to impose.
> There are several literals in that code snippet; I could annotate them with their types, and this code would still be exactly as it is.
Well, no, there is one literal that has an ambiguous type, and if you annotated its type, it would resolve entirely the question of what a typechecker should say; literally the entire reason it is an open question is because that one literal is not annotated.
True, you could annotate 3 of the 4 literals in this without annotating the List, which is ambiguous. In the absence of an explicit annotation (because those are optional), type checkers are left to guess intent to determine whether you wanted a List[Any] or List[number | string], or whether you wanted a List[number] or List[string].
Right. And the fact that python doesn't specify the semantics of its type annotations is a super interesting experiment.
Optimally, this will result in a democratic consensus of semantics.
Pessimistically, this will result in dialects of semantics that result in dialects of runtime languages as folks adopt type checkers.