Parse Python Function's Class Name

Question

I am trying to find the class name where a function name is called.

E.g. suppose I have:

class foo1(object):
    bar1()

class foo2(object):
    bar2()

class foo3(object):
    bar3()

If I am searching for bar2() I want to receive:

class foo2(object):
    bar2()

I tried regex class[\s\S]*bar2\(\)(?!class) with the idea being that I would negative lookahead for another appearance of class. Unfortunately it looks like the [\s\S] is already matching everything: https://regex101.com/r/kZ7eE5/1

Is there a way to match if "class" appears only one time while matching on all other chars (including new line and tab)?

Also alternatives that do not require regex would be accepted.

Your chances might be better just scanning one line after the other and memorizing the last class definition you encountered. Also, what about nested classes, or methods, or methods with nested classes? — tobias_k
– tobias_k, Commented Jun 26, 2015 at 15:30
I want to analyze the code base. Is there another tool that I can parse Python with? — Alan Liang
– Alan Liang, Commented Jun 26, 2015 at 15:31
Not used it myself, but docs.python.org/2/library/parser.html might be worth a look, or the ast(docs.python.org/2/library/ast.html) module. — tobias_k
– tobias_k, Commented Jun 26, 2015 at 15:33
@tobias_k yea I'll try that. Good point about the nested classes - in that case I would want just the closest level class. If the method is in a method which is in a class, I would want that class. — Alan Liang
– Alan Liang, Commented Jun 26, 2015 at 15:37

fferri · Accepted Answer · 2015-06-26 19:06:02Z

5

The RE approach can be error prone (the expressive power of Python language is more that those of the regular languages recognized by REs).

Use Python's ast module to parse Python code:

code = '''
class foo1(object):
    bar1()

class foo2(object):
    bar2()

class foo3(object):
    bar3()
'''

>>> import ast
>>> tree = ast.parse(code)
>>> for i in tree.body:
...     if isinstance(i, ast.ClassDef):
...         print('class: %s' % i.name)
...
class: foo1
class: foo2
class: foo3
>>>

and you can do many other things, check the docs at https://greentreesnakes.readthedocs.org/en/latest/

Edit: a more complete example:

>>> for i in tree.body:
...     if isinstance(i, ast.ClassDef):
...         for j in i.body:
...             if isinstance(j, ast.Expr):
...                 print('found a call to function %s in class %s' % (j.value.func.id, i.name))
...
found a call to function bar1 in class foo1
found a call to function bar2 in class foo2
found a call to function bar3 in class foo3

edited Jun 26, 2015 at 19:06

answered Jun 26, 2015 at 15:36

fferri

19k8 gold badges57 silver badges110 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Pruthvi Raj Over a year ago

This doesn't serve OPs purpose, he wants to return a class that contains a certain function call

tobias_k Over a year ago

@PruthviRaj It's not complete, but it certainly works better than using regex.

Pruthvi Raj Over a year ago

@mescalinum , +1 this is perfect :)

Pruthvi Raj · Accepted Answer · 2015-06-26 15:35:12Z

1

The following code prints the class that contains the method call that is stored in function variable .Here's what you need to do :

import re

data = """
class foo1(object):
    bar1()

class foo2(object):
    bar2()

class foo3(object):
    bar3()
"""

classes = re.findall('(class[\s\S]*?)(?=class|$)',data)
function = "bar2()"
for x in classes:
    if function in x:
        matched_class = x
        break
print(matched_class)

Output:

class foo2(object):
    bar2()

answered Jun 26, 2015 at 15:35

Pruthvi Raj

3,0363 gold badges25 silver badges36 bronze badges

4 Comments

tobias_k Over a year ago

Now try to use this to find a method named get_class() or similar.

Pruthvi Raj Over a year ago

@tobias_k , I agree this approach might not be better but I did what OP has asked for

tobias_k Over a year ago

He did not ask for this specifically, just wanted to point out that the regex appraoch (not just yours, but generally) is not really up to the job. That's just one example of when it will fail.

Pruthvi Raj Over a year ago

yeah I see the approach by mescalinium is better :) , Thank you

melwil · Accepted Answer · 2015-06-26 15:32:24Z

0

The reason [\s\S]* matches everything is because you are telling is to match either any space character (\s) or any non-space character (\S) any number of times. You can just match space with space in regex. Usually. There are exceptions, but python demands this form, so it should be fine.

You can use

^class ([\w]+)\([^)]+\):

Demo: https://regex101.com/r/aB9pX4/2

answered Jun 26, 2015 at 15:32

melwil

2,5531 gold badge21 silver badges34 bronze badges

1 Comment

Alan Liang Over a year ago

This will return all classes, but what if I want only the class with function bar2()?

Collectives™ on Stack Overflow

Parse Python Function's Class Name

3 Answers 3

3 Comments

4 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

4 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related