Restrict python function to take only list as an argument

Question

Is it possible in python to restrict parameter type in python. I tried using List module from typing, but it still allows user to pass class object or an int parameter without giving error.

 from typing import List

 class Job:
    def __init__(self):
            self.profit = 0
            self.deadlines = 0

 class JobsWithDeadlines:
    def get_Job_Sequencing_with_deadlines(self, jobs: List[Job]):
            print(type(jobs))


 j = Job()
 j.profit = 20
 j.deadline = 2
 obj = JobsWithDeadlines()
 obj.get_Job_Sequencing_with_deadlines(1)

score 1 · Answer 1 · answered Sep 06 '19 at 03:47

1

While type annotations are a part of Python, type checking is not. The annotations serve to enable external static type checking. The most common static type checkers for Python are the PyCharm editor and mypy.

"static" means catching the errors before the program is run, just by analysing the source code. There is nothing that can prevent a wrong object being passed at runtime, except an explicit check.

answered Sep 06 '19 at 03:47

Amadan

191,408
23
240
301

can you share a code example, because I see mypy library to be using same typing module that i have used. Actually , I am not able to use mypy in my code – Rahul Ranjan Sep 06 '19 at 04:26
1

There is no code example. You are not supposed to use mypy _in_ your code - it is a _static_ code analyser, as I explained in the answer. `mypy my_program.py` (from command line) will check your code. It will complain if it is _possible_ (given the code it sees) that your function would receive a wrongly-typed argument; then if you eliminate all the warnings from the static analyser, runtime checks are not needed. – Amadan Sep 06 '19 at 04:27

Mad Physicist · Answer 2 · 2019-09-06T05:25:23.807

The notation jobs:List[Job] is called annotation. Annotations provide a way of writing hints for the user and for IDEs. Annotations are stored in the function object at runtime, but do not impose any sort of enforceable restriction on the actual code.

To enforce the restriction, you have to check it manually. For example, if subclasses are allowed:

def get_Job_Sequencing_with_deadlines(self, jobs):
    if not isinstance(jobs, list):
         raise TypeError('non-list input')
    print(type(jobs))

If you don't want to allow subclasses, use if type(jobs) is not list: instead.

Checking each item in a list is equally painful:

if not all(isinstance(job, Job) for job in jobs):
    raise TypeError('non-Job input')

You almost never want to do explicit type checking in Python. The language is fundamentally designed with the idea of duck typing in mind. This means that objects that behave the way you want but aren't part of some hard coded class hierarchy should be allowed. Pragmatically this makes perfect sense. If it behaves like a list, do you care if it isn't?

Restrict python function to take only list as an argument

2 Answers2