How to use new stricter definition of pandas DataFrame.itertuples()?

[j-carson]

With the most recent pandas-stubs, mypy now complains about every call to itertuples()
Is there a way that I'm supposed to tell the new release what my column names are?

Sample

import pandas as pd                                                                                 
                                                                                                    
df = pd.DataFrame({"a": [1, 2, 3, 4], "b": [5, 6, 7, 8]})                                           
                                                                                                    
for item in df.itertuples():                                                                        
    a = item.a                                                                                      
    b = item.b     

mypy.....................................................................Failed
- hook id: mypy
- exit code: 1

foo.py:6: error: "tuple[Any, ...]" has no attribute "a"  [attr-defined]
foo.py:7: error: "tuple[Any, ...]" has no attribute "b"  [attr-defined]
Found 2 errors in 1 file (checked 1 source file)

[twoertwein]

It would probably be best to return Iterable[Any] to cover any possible field of a namedtuple. #822 assumed that itertuples returns a normal tuple.

[Dr-Irv]

Prior to #822 , we had no return type. I think Iterable[Any] is the best solution.

[Dr-Irv]

Created #834