Well-featured retry decorator
From the Jargon File:
The -P Convention Turning a word into a question by appending the syllable ‘P’; from the LISP convention of appending the letter ‘P’ to denote a predicate (a boolean-valued function). The question should expect a yes/no answer, though it needn't. (See T and NIL.)
retryp is yet another retry decorator, with nothing particularly special about it other than it being both well-featured and not rewriting or hiding the calling signature of the wrapped method (allows for accurate code introspection – a Big Deal for me).
Thanks go to Graham Dumpleton and his wrapt package for that latter.
A multiplicative factor applied to the delay, used to provide increasing backoff as subsequent attempts continue to fail.
Number of times to retry the wapped item. If set to “0”, will retry 1073741823 times because a) that’s a lot, b) give up already, and c) nothing lasts forever.
Default: 10 (seconds)
The base delay between retry attempts.
By default retryp will raise retryp.FailedTooOften if the wrapped item continues to fail after the requested number of attempts. If expose_last_exc is set, then the exception raised by the wrapped item will be raised if the last attempt results in an exception.
Extra random seconds will be added to each retry delay, ranging from 0 to the value of jitter.
Log every exception raised by the wrapped item using logtool.log_fault.
Logging level at which exceptions will be logged by logtool.log_fault() when log_faults is set.
String to use in log messages to describe what is being retried.
If provided, the return of the wrapped item will be passed as the only argument to this function. If refuse_rc_fn (rc) evaluates to True, then no further retries will be made.
If provided, any exception raised by the wrapped item will be passed as the only argument to this function. If refuse_exc_fn (e) evaluates to True, then the exception will be raised again and retry attempts will cease.