How to fold python code using sed and/or awk?

November 2, 2024

I have a large python code base. I want to get a feel of how a BaseClass is subclassed by ‘grepping-out’ the name of the sub class and the functions in the class, but only for classes that inherit from SomeBaseClass.

So if we have multiple files that have multiple classes, and some of the classes look like:

class SubClass_A(BaseClass):
    def foo(self):
        ...

    async def goo(self):
        ...

class AnotherClass:
    ...

class SubClass_B(BaseClass):
    def foo(self):
        ...

    async def goo(self):
        ...

If we apply some sort of scrip to this code, it would output:

File: /path/to/file
class SubClass_A(BaseClass):
    def foo(self):
    async def goo(self):
class SubClass_B(BaseClass):
    def foo(self):
    async def goo(self):

There can be many classes in the same file some of them may be a BaseClass. Essentially, this is like folding lines of code in an IDE.

Now, I tried using sed to do this, but I’m not an expert in it, and sed won’t be able to print the file path. However, I know awk can print the file path, but I don’t know awk! Argh.

I thought about writing a python program to do this, but this problem seems like something a nifty sed/awk program can do.

Thanks for the help.

>Solution :

Using any awk, this may be what you’re trying to do:

$ awk '/^class/{ f=/\(BaseClass)/ } f && $1 ~ /^(class|def|async)$/' file
class SubClass_A(BaseClass):
    def foo(self):
    async def goo(self):
class SubClass_B(BaseClass):
    def foo(self):
    async def goo(self):