mypy provides optional type-checking for python code. See the docs of mypy.

After installing it with pip install mypy you should be able to run

mypy  some_file.py

or maybe

mypy --ignore-missing-imports some_directory/

to avoid certain warnings.

mypy will already find an warn about some issues even without you doint any extra work.

mypy success

Before showing all the problems mypy can find let's spend a few seconds to see what happens if our code does not have any issues:

examples/python/mypy/hw.py

print("Hello World")

$ mypy hw.py
Success: no issues found in 1 source file

mypy without annotation

examples/python/mypy/a.py

a = 23
a = "hello"

This code does not have any annotation yet, but we put different values in the same variable and mypy will report it.

$ mypy a.py
a.py:2: error: Incompatible types in assignment (expression has type "str", variable has type "int")
Found 1 error in 1 file (checked 1 source file)

A more interesting case might be this one:

examples/python/mypy/aa.py

n = input("A number: ")
n = int(n)

We expect the user to type in an integer and we convert it to an integer. We need to convert it, because input always return a string. So in this rather common code we can see how mypy suggest that we write stricter code.

$ mypy aa.py
aa.py:2: error: Incompatible types in assignment (expression has type "int", variable has type "str")
Found 1 error in 1 file (checked 1 source file)

One of the solutions is to use two separate variable names:

examples/python/mypy/aaa.py

n_as_str = input("A number: ")
n = int(n_as_str)

$ mypy aaa.py
Success: no issues found in 1 source file

mypy and functions without annotation

If we have similar isse but inside a function like here:

examples/python/mypy/b.py

def f():
    x = 23
    x = "hello"

f()

mypy will disregard this. By default mypy does not look inside functions.

$ mypy b.py
Success: no issues found in 1 source file

I have not found a way to tell mypy to look into those functions without changing my code. If we add the smallest annotation to a function, mypy will start checking the whole function.

So here I declared that the function returns None.

examples/python/mypy/c.py

def f() -> None:
    x = 23
    x = "hello"

f()

$ mypy c.py
c.py:4: error: Incompatible types in assignment (expression has type "str", variable has type "int")
Found 1 error in 1 file (checked 1 source file)

mypy catching a bug

Let's see an example, still a simplified one, in which mypy will help us discover a bug. If course writing extensive unittest would uncover such bugs, but if the code is complex (unlike this example) then it is quite likely you won't be able to cover all the cases.

Moreover this shows that even if you don't have time to add unittest, adding simple annotation to your code and using mypy will already allow you to find problems.

examples/python/mypy/reg.py

import re

def check(regex, text) -> None:
    match = re.search(regex, text)
    print(match)
    z = match.group(0)
    print(z)

check('cat', 'Wilde cat')
check('dog', 'Wilde cat')
print('done')

As long as there Is a match the above code works, but if there is no match this python code raises an exception.

Using mypy we can find out about this problem as mypy gives the following error:

$ mypy reg.py
reg.py:7: error: Item "None" of "Optional[Match[str]]" has no attribute "group"
Found 1 error in 1 file (checked 1 source file)

Conclusion

These example can already show you the usefulness of mypy. In upcoming articles we'll see more use-cases.