Decorator for retrying a function N times by catching one of the specified exceptions and then retrying. Specially useful for functions that throws errors sporadically, like ones that depends on external resources as web APIs, databases, etc.

Installation:

pip install retry_on_exceptions

or through source code:

git clone git://github.com/fjsj/retry_on_exceptions.git
cd retry_on_exceptions
python setup.py install

Usage:

from retry import retry_on_exceptions

current_try = 0
@retry_on_exceptions(types=[ZeroDivisionError, KeyError], tries=3)
def test():
    global current_try
    current_try += 1
    if current_try == 1:
        return 1 / 0
    elif current_try == 2:
        return dict()['key']
    else:
        return "Got it on last try!"    

if __name__ == "__main__":
    print test()

The code above prints (if root logger is active and on debug level):

• Retrying function test
• Retrying function test
• Last try... and I will raise up whatever exception is raised
• Got it on last try!

Optionally, you can also specify a delay (a float in seconds), making the current thread sleep between tries:

@retry_on_exceptions(types=[urllib2.URLError], tries=3, delay=3.5)