How to read and strip only specific lines of code?

Question

I am writing a python program which reads C++ program files from a folder. The program then strips only certain parts from the C++ code read and writes them into a new separate file.

This is the C++ code that I am trying to strip:

#include <stdio.h>

int max(int num1, int num2);

//selectmethod 
int main () {
   int a = 100;
   int b = 200;
   int ret;
   ret = max(a, b);
   printf( "Max value is : %d\n", ret );
   return 0;
}

int max(int num1, int num2) {
   int result;
   if (num1 > num2)
      result = num1;
   else
      result = num2;
   return result; 
}

The only parts I am interesting in taking are marked with a //selectmethod comment. In the example code above, this would be the main() method.

Right now I have a python code which reads this file and writes all of the file contents into an output file. However, I want to modify my code so the output file will only contain this:

int main () {
   int a = 100;
   int b = 200;
   int ret;
   ret = max(a, b);
   printf( "Max value is : %d\n", ret );
   return 0;
}

The python code I have so far is below:

import glob
import os.path

list_of_files = glob.glob('/my/input/files/*.cc')

def main():
    for file_name in list_of_files:
        print(file_name)

        f= open(file_name, 'r')
        lst = [];
        plist = [];
        for line in f:
           fline = line.strip()
           lst.append(fline)
           plist.append(line)
        f.close()

        print(lst)

        f=open(os.path.join('/my/output/files/path',
        os.path.basename(file_name)), 'w')

        for line2 in plist:
           f.write(line2)
        f.close()


if __name__ == "__main__":
    main()

How can I modify my code to only extract the method below lines starting with //selectmethod?

What does this have to do with NLP? C++ is not a natural language, it's a programming language. — Barmar, Nov 19 '19 at 01:05
Are you asking how to write a C++ parser in Python, so it can tell where a function begins and ends? That's not a simple task. — Barmar, Nov 19 '19 at 01:06
If you can assume that the code is properly indented, you can simply copy everything from `//selectmethod` to the next line that contains just `}`. — Barmar, Nov 19 '19 at 01:07
@Barmar, what is there is an `if` or `for` in the loop? It will also have a `}`. — Harshal Parekh, Nov 19 '19 at 01:11
In this case, you will have to keep a track of how many `{` were encountered. — Harshal Parekh, Nov 19 '19 at 01:12
But the `}` won't be at the beginning of the line, there will be spaces before it. That's what I meant by "assume that the code is properly indented". — Barmar, Nov 19 '19 at 01:13
@Barmar yes, the assumption should be made that the code should be properly indented — user5455438, Nov 19 '19 at 01:14
@HarshalParekh yes, it should also accomodate if and else statements — user5455438, Nov 19 '19 at 01:15
As @Barmar asked, can we assume the code in properly indented? — Harshal Parekh, Nov 19 '19 at 01:16
@HarshalParekh yes we can assume that the code is properly indented — user5455438, Nov 19 '19 at 01:16
And the comment `//selectmethod`, is it there in the code, or you put it explicitly? — Harshal Parekh, Nov 19 '19 at 01:17
C++ code may contain raw string literals (that may contain single `}` on a line) and namespaces (that are usually also end with single `}` on a line). So in general case this task is equal to creating C++ parser even if code is "properly" formatted :) — 4LegsDrivenCat, Nov 19 '19 at 01:29

score 1 · Answer 1 · answered Nov 19 '19 at 01:35

1

You can do it using a simple Regex expression, here's a sample code below When searching, the expressions uses the single line option re.S to treat the file as a single string, allowing to capture multi lines. Regex is a very powerful and efficient way to search/replace text, for more information see https://www.w3schools.com/python/python_regex.asp

import re

#a multi line string representing the C file
string = """
#include <stdio.h>

int max(int num1, int num2);

//selectmethod 
int main () {
   int a = 100;
   int b = 200;
   int ret;
   ret = max(a, b);
   printf( "Max value is : %d\n", ret );
   return 0;
}

int max(int num1, int num2) {
   int result;
   if (num1 > num2)
      result = num1;
   else
      result = num2;
   return result; 
}
"""

result = re.search("int main\s*\(\s*\)\s*\{.*?\}",string, re.S).group()
print(result)

answered Nov 19 '19 at 01:35

Mahmoud Hanafy

703
9
11

This is the sort of solution that I am looking to implement. Would this still work for different types of methods (for example methods that are not "main" and methods that have lot of if and else statements inside them – user5455438 Nov 19 '19 at 02:06
This can be done using recursion or subroutines, i.e. this expression will match exactly what you want `int\s*main\s*(\{(?:[^}{]*|(?1))*\})` – Mahmoud Hanafy Nov 19 '19 at 03:58
The above is using subroutines, which isn't natively supported by Python; this one uses recursion in python but can't match the `int main` part `\{[^}{]*+(?:(?R)[^}{]*)*+\}` – Mahmoud Hanafy Nov 19 '19 at 03:59
credit to this answer here https://stackoverflow.com/a/35271017/2136590 – Mahmoud Hanafy Nov 19 '19 at 04:01

4LegsDrivenCat · Answer 2 · 2019-11-19T14:41:11.313

In general, this task is equal to creating full C++ parser even if code is "properly" formatted. For those who look for single } character on a line, here as an example of C++ code that contains two false positives in a raw string literal:

#include <iostream>
int main()
{
    std::cout << R"(Rules: You may use any JSON string but the following three characters are forbidden by Big Brother Inc.:
$
!
}

Example of JSON string:
{
  "name":"value"
}
)";
    return 0;
}

C++ code may also contain namespaces that are usually end with single } on a line.

Although you may be lucky and your C++ code is quite simple so you don't need full C++ parser :)

If someone marked the beginning of functions of interest by //selectmethod marker, ask that person to also mark the end of those functions by some other marker :)

How to read and strip only specific lines of code?

2 Answers2