Pandas: how to pass multiple column values into apply function?

I have an task where I have to find a weight value in dataframe, it can be found in several columns (but not every column though)

Simplified dataframe looks like this:

          str1         str2
0  some string         2kg.
1      got 5kg  some string
2  some string  some string

Simplified example of how it works right now:

import re
import pandas as pd
from pandas import Series

regex = re.compile(r'(d+)kg')

def find_weight(row: Series) -> int | None:
    # first check in 'str1'
    result = re.search(pattern=regex, string=row['str1'])
    if result:
        return int(result.group(1))
    # if not found, check 'str2'
    result = re.search(pattern=regex, string=row['str2'])
    if result:
        return int(result.group(1))
    # if nothing found in both strings - return None
    return None


df = pd.DataFrame([['some string', '2kg.'], ['got 5kg', 'some string'], ['some string', 'some string']], columns=['str1', 'str2'])
df['weight'] = df.apply(find_weight, axis=1)
print(df)

However, in a real-life case, I have 40+ columns and weight value can be found only in 2 of them. Passing an entire row into a function every time feels wasteful. How can I pass only arguments from several columns instead of the entire row?

How to pass values only of colmns ‘str1’ and ‘str2’ into function and then apply it to every row?