Regex Pattern to find arguments passed into a function

Question

For a project, I am trying to read through a python file and keep a list of all the variable being used within a certain function. I am reading through the lines in the python file in string format and then focusing on a line where starting with "def". For the purpose of this example pretend we have the following line identified:

def func(int_var:int,float_var=12.1,string_var=foo()):

I want to use regex or any other method to grab the values within this function declaration.

I want to grab the string "int_var:int,float_var=12.1,string_var=foo()", and later split it based on the commas to get ["int_var:int","float_var=12.1","string_var=foo()"]

I am having a lot of trouble being able to isolate the items between the parenthesis corresponding to 'func'.

Any help creating a regex pattern would be greatly appreciated!

• Xavier C: I'd also like to grab the function name ('func') as a separate variable!

Answer 1

Instead of regex, it is much easier and far more robust to use the ast module:

import ast
s = """
def func(int_var:int,float_var=12.1,string_var=foo()):
 pass
"""
def form_sig(sig):
 a = sig.args
 d = [f'{ast.unparse(a.pop())}={ast.unparse(j)}' for j in sig.defaults[::-1]][::-1]
 v_arg = [] if sig.vararg is None else [f'*{sig.vararg.arg}']
 kwarg = [] if sig.vararg is None else [f'*{sig.kwark.arg}']
 return [*map(ast.unparse, a), *d, *v_arg, *kwarg]
f = [{'name':i.name, 'sig':form_sig(i.args)} for i in ast.walk(ast.parse(s)) 
 if isinstance(i, ast.FunctionDef)] 

Output:

[{'name': 'func', 'sig': ['int_var: int', 'float_var=12.1', 'string_var=foo()']}]

• Xavier C: Thank you for taking the time to write this up! Super useful.

Answer 2

func_pattern = re.compile(r'^\s*def\s(?P<name>[A-z_][A-z0-9_]+)\((?P<args>.*)\):$')
match = func_pattern.match('def my_func(arg1, arg2):')
func_name = match.group('name') # my_func
func_args = match.group('args').split(',') # ['arg1', 'arg2']

• Xavier C: I appreciate this!
• Xavier C: I noticed that this solution fails when there are spaces before or after the commas. For example: This works --> def foo(x,y,z=foo1()) This fails --> def foo( x ,y , z=foo1()) I was trying to see the pattern for identifying the arguments but couldn't determine how to modify it to allow for spaces.
• Ethan Olpin: I'm not sure I understand. The ".*" found in the pattern should match any type of character in between the parentheses, spaces included. A quick check on my end indicates that the pattern matches on both of the examples you gave (after adding a colon to the end, strings not ending with a colon will not be matched by the pattern I gave). Perhaps the issue is that we don't want spaces in the output? If you would like to remove the whitespace that show up around the arguments after you get your match you could do: func_args = [arg.strip() for arg in match.group('args'),split(',')]
• Anina7: The split(',') does not work in cases of argument types like "dict[str, int]"