# Example: Functions argument types#

In this example, we are going to see what types of arguments are most popular in Python.

First, let’s pull in all the code from every module we can import:

from code_data.module_codes import modules_codes_cached

# tuples of name, source, code
module_codes = modules_codes_cached()

We have loaded 5795 modules


Now we can turn them all into data, and also iterate through them all to have all the recursive ones at the top level:

from code_data import CodeData

all_code_data = {code_data for _, _, code in module_codes for code_data in CodeData.from_code(code).all_code_data()}
print(f"We have loaded {len(all_code_data)} code data objects")

We have loaded 155774 code data objects


Let’s filter for functions:

from code_data import Function

fns = {c for c in all_code_data if isinstance(c.type, Function)}
print(f"{len(fns)} of them are functions")

127087 of them are functions


We can see how many functions have a docstring:

from collections import Counter

have_docstring = Counter(bool(f.type.docstring) for f in fns)

print(f"The function has a docstring? {have_docstring}")

The function has a docstring? Counter({False: 96609, True: 30478})


And finally, we can see how many different argument types are used cumulatively:

param_kinds = Counter(k for c in fns for k in c.type.args.parameters.values())
param_kinds

Counter({<_ParameterKind.POSITIONAL_OR_KEYWORD: 1>: 203284,
<_ParameterKind.VAR_KEYWORD: 4>: 5067,
<_ParameterKind.VAR_POSITIONAL: 2>: 3701,
<_ParameterKind.KEYWORD_ONLY: 3>: 3286,
<_ParameterKind.POSITIONAL_ONLY: 0>: 246})


There are an order of magnitude more “positional or keyword” arguments then all other types combined!

And we can see there is the least number of positional only arguments, which make sense since they were most recently introduced in PEP 570