laike9m / python-type-challenges Goto Github PK

this one felt a bit too easy -- I'm not sure how to change the problem to ...

from typing import *


class Fn[R, **P]:
    def __init__(self, f: Callable[P, R]):
        self.f = f

    def transform_callable(self) -> Callable[Concatenate[object, P], R]:
        ...

As we have random challenge entrance, we can add it to README.

Example

https://github.com/laike9m/Python-Type-Challenges/blob/main/challenges/advanced-overload-literal/solution.py

Currently, they have to click the x button which is annoying an counterintuitive

It is also necessary to explicitly specify the Python version in the context, otherwise a lot of the solving will be affected.

this works locally -- and on the version of pyright stated in requirements.txt

from typing import Protocol


class Person:
    pass


class P1[R, **P](Protocol):
    def __call__(self, value: object, *args: P.args, **kwargs: P.kwargs) -> R: ...

class P2[R, **P](Protocol):
    def __call__(self, value: Person, *args: P.args, **kwargs: P.kwargs) -> R: ...

def transform[R, **P](f: P1[R, P]) -> P2[R, P]:
    def wrapper(value: Person, *args: P.args, **kwargs: P.kwargs) -> R:
        return f(value, *args, **kwargs)

    return wrapper

peeked at the stated solution and it feels a little cheeky to rely on the type checker to fill in the return value there

Python version will affect the challenge test, such as:

Challenge - dict

type custom = dict[str, str]

def foo(x: custom):
    pass

it's ok in python 3.12, but could not pass the test.

I think it related with #92

PEP 692

Example:

It is a bit difficult to locate which line is incorrect.

Maybe we can split the challenge into source code part and test part.

It would be nice to show the status of each challenges (passed, not passed), like Leetcode:

We can store the information in localStorage

Currently, there is no hint message for most questions. For a few questions like callable-protocol, there is a short "HINT: Use Protocol" written in the code, but there is no standard.

We do provide a "Solution" button, but it's very different from a hint. When the user is playing with each challenge, a well-designed hint message can guide him and help him complete the question by learning, but the "solution" button takes him directly to the answer--It's faster but the user may not learn anything at all!

So I'm suggesting that we add a hint message to all questions. The hint messages follow a standard format, like this:

• For basic/any: HINT: Check out Any.
• For intermediate/callable: HINT: Check out Callable.
• For intermediate/generic2: HINT: Check out TypeVar, constraints are needed for this question.
• ...

Implementation

The hint can be added directly in any question.py like this:

## HINT start ##
(message in Markdown format)
## HINT end ##

The server can extracts the message and render it as HTML(to make hyper link works).

something like

basic       intermediate      advanced     extreme
basic-1    intermediate-1    advanced-1   extreme-1
basic-2    intermediate-2    advanced-2   extreme-2
   ...

Right now it's a single column, which doesn't look good

For implementation, we won't be reusing the whole challenge_list component, but we perhaps can make a component to represent challenges belonging to a single level, and reuse it.

Test should cover returning challenges from backend, and sending user code to backend and get type check result

@100gle How did you get the Python svg logo? I assume you used https://icon-sets.iconify.design/logos/python/ ?

Right now there are 4 levels

• basic
• intermediate
• advanced
• extreme

Though there's no predefined criterias for determing a challenge's level.

It seems pretty reproducible, but I'm not sure if this is our problem or theirs

The following fails to work on the typealias challenge:

"""
TODO:

Create a new type called Vector, which is a list of float.
"""

type Vector = list[float]

with the error:

Error:
7:error: Statements must be separated by newlines or semicolons

This took me a second, and the solution that's accepted is in fact:

"""
TODO:

Create a new type called Vector, which is a list of float.
"""

type Vector = list[float]
 

Note the trailing newline (I had to add a space as well because the GH markdown parser seems to swallow it otherwise).
I haven't looked at the code, but I suspect it's concatenating the left and right code samples without adding a newline, so this could come as a surprise to some.

It would in general maybe be better to just strip both the LHS and the RHS of any trailing whitespace, and concatenate them with a newline in the middle, just to avoid issues with indentation and newlines.

PEP 646

Description

def f(a: list[int | str]):
    pass


def g(a: str | list):
    pass

Suggestion

Maybe we can use definition to make challenge strict.

pyright supports switch python version with command line option --pythonversion

Currently the challenges appear in alphabetical order, it would be better if we customize the order by prioritizing the easier challenges to the top. For example, the basic challenges could use this order:

# The basic syntax
- parameter
- return
- variable
- any
- kwargs
# Composite data types
- list
- dict
- tuple
# Other
- typealias
- union
- optional
- final

This would make the experience better as people go through the challenges from top to bottom.

solution.py in recursive uses TypeAlias which is deprecated since 3.12.

Image size should be 1200 x 630 to fit the requirement of social networks

dict[str: str] does not seem to be correct, and it should be dict[str, str].

I imagine that imports were left out to avoid giving clues, but it would be nice to have all the type hints imported into the namespace.

Zeabur free plan has a memory limit of 512MB, and it appears that we're running on the edge and (according to their dev team) have OOM'ed

PEP 696 defaults for type variables

Description

After pdm dev, UnicodeDecodeError(gbk: illegal multibyte sequence) caused by string ——

the string is in Python-Type-Challenges/challenges/intermediate-self/hints.md

Also, the emoji cannot load successfully

Solution

Python UTF-8 mode

Scenario:
I write some codes, it does not work so I want to start over again. I want to remove all my code quickly and start from the original codes.

PEP 698

In your solution to the extreme concatenate problem you propose:

def transform[T, **P](f: Callable[Concatenate[Any, P], T]):
    def wrapper(value: Person, *args: P.args, **kwargs: P.kwargs) -> T:
        return f(value, *args, **kwargs)

    return wrapper

But if I add the return type of Callable[Concatenate[Person, P], T] to the transform function, the test in line 19 fails.

Is there a better return type or do you recommend skipping it in general?

Description

When I click the button online, the server return a 500 error.

However, I can run locally, it seems not error for python.
Or 84551ad cause?

In challenge intermediate/literalstring, function execute_query has a parament accepted Iterable[str]，and set default value .... The same pattern is observed in the provided solution.

def execute_query(sql, parameters: Iterable[str] = ...):
    """No need to implement it"""

This will cause a type check error by pyright

error: Expression of type "ellipsis" cannot be assigned to parameter of type "Iterable[str]"

It seems like I must access type check by changing Iterable[str] to Iterable[str] | EllipsisType , like

from typing import Iterable, LiteralString
from types import EllipsisType


def execute_query(sql: LiteralString, parameters: Iterable[str] | EllipsisType = ...):
    """No need to implement it"""

or change ... to () for type check.

from typing import Iterable, LiteralString
from types import EllipsisType


def execute_query(sql: LiteralString, parameters: Iterable[str] = ()):
    """No need to implement it"""

I don't think either of the above solutions to be the best way for this issue, but I don't know any better way to use ellipsis for default parameter with type hint.

Hello,
I was trying to solve the last challenge.
this one

but I was getting strange errors in the output, so
I decided to test out the real solution but again, pasting & running the solution produced errors:

Description

In black 24.1.0, many challenges will be formatted by black

Dummy class and function implementations consisting only of ... are formatted more
compactly

class Undergraduate:
    ...

It will be replaced with

class Undergraduate: ...

In solution.py, user should complete ...

class Undergraduate(Student):
    major: str

Suggestion

Replace black with ruff

Bug description

If # expect-type-error not in test_code, challenge can pass with raise.

Challenges

• advanced-forward
• advanced-never2

In case people accidentally changed on but forget about the other

PEP 675

https://github.com/RF-Tar-Railt/Python-Type-Challenges/blob/main/challenges/advanced-overload-generic/question.py

https://github.com/RF-Tar-Railt/Python-Type-Challenges/blob/main/challenges/extreme-constructor-parameter/question.py

mypy

pyright only reports errors in test code, which is what we expected

This issue should be fixed once we migrate to pyright (#9)

cc @RF-Tar-Railt @frostming

Description

Currently, click random button will reload the page, and all passed-icon will reload.