Skip to main content

Exit programs gracefully.

Project description

pyterminate

CI PyPI Downloads

Reliably run cleanup code upon program termination.

Table of Contents

Why does this exist?

There are currently two builtin modules for handling termination behavior in Python: atexit and signal. However, using them directly leads to a lot of repeated boilerplate code, and some non-obvious behaviors that can be easy to accidentally get wrong, which is why I wrote this package.

The atexit module is currently insufficient since it fails to handle signals. The signal module is currently insufficient since it fails to handle normal or exception-caused exits.

Typical approaches would include frequently repeated code registering a function both with atexit and on desired signals. However, extra care sometimes needs to be taken to ensure the function doesn't run twice (or is idempotent), and that a previously registered signal handler gets called.

What can it do?

This packages does or allows the following behavior:

  • Register a function to be called on program termination

    • Always on normal or exception-caused termination: @pyterminate.register
    • Configurable for any desired signals:
      @pyterminate.register(signals=(signal.SIGINT, signal.SIGABRT))
  • Allows multiple functions to be registered

  • Will call previous registered signal handlers

  • Allows zero or non-zero exit codes on captured signals:
    @pyterminate.register(successful_exit=True)

  • Allows suppressing or throwing of KeyboardInterrupt on SIGINT:
    @pyterminate.register(keyboard_interrupt_on_sigint=True)

    • You may want to throw a KeyboardInterrupt if there is additional exception handling defined.
  • Allows functions to be unregistered: pyterminate.unregister(func)

  • Ignore requested signals while registered function is executing, ensuring that it is not interrupted.

    • It's important to note that SIGKILL and calls to os._exit() cannot be ignored.

Quickstart

python3 -m pip install pyterminate
import signal

import pyterminate


@pyterminate.register(
    args=(None,),
    kwargs={"b": 42},
    signals=(signal.SIGINT, signal.SIGTERM),
    successful_exit=True,
    keyboard_interrupt_on_sigint=True
)
def cleanup(*args, **kwargs):
    ...

# or

pyterminate.register(cleanup, ...)

Tips, tricks, and other notes

Duplicate registration after forking

Since creating a new process through forking duplicates the entire process, any previously registered functions will also be registered in the forked process. This is an obvious consequence of forking, but important to consider if the registered functions are accessing shared resources. To avoid this behavior, you can unregister the function at the beginning of the forked process, gate based on the process' ID, or use any other synchronization method that's appropriate.

Multiprocessing start method

When starting processes with Python's multiprocessing module, the fork method will fail to call registered functions on exit, since the process is ended with os._exit() internally, which bypasses all cleanup and immediately kills the process.

One way of getting around this are using the "spawn" start method if that is acceptable for your application. Another method is to register your function to a user-defined signal, and wrap your process code in try-except block, raising the user-defined signal at the end. pyterminate provides this functionality in the form of the exit_with_signal decorator, which simply wraps the decorated function in a try-finally block, and raises the given signal. Example usage:

import multiprocessing as mp
import signal

import pyterminate


@pyterminate.exit_with_signal(signal.SIGUSR1)
def run_process():

    @pyterminate.register(signals=[signal.SIGUSR1, signal.SIGINT, signal.SIGTERM])
    def cleanup():
        ...

    ...


if __name__ == "__main__"
    mp.set_start_method("fork")

    proc = mp.Process(target=run_process)
    proc.start()

    try:
        proc.join(timeout=300)
    except TimeoutError:
        proc.terminate()
        proc.join()

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

pyterminate-0.0.8.tar.gz (6.2 kB view details)

Uploaded Source

Built Distribution

pyterminate-0.0.8-py3-none-any.whl (6.5 kB view details)

Uploaded Python 3

File details

Details for the file pyterminate-0.0.8.tar.gz.

File metadata

  • Download URL: pyterminate-0.0.8.tar.gz
  • Upload date:
  • Size: 6.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.9.17

File hashes

Hashes for pyterminate-0.0.8.tar.gz
Algorithm Hash digest
SHA256 31f1cead91cd425eda642f16d20e5ea0ce35579caf89a30cf1992bf3d418457b
MD5 7755c5b0548fa0b7ee16675cb2f7dd8a
BLAKE2b-256 b445d65ea35eb77a8121593b4a6a46fdac9467f1eb5f5d6b83c0b9295cc738d5

See more details on using hashes here.

File details

Details for the file pyterminate-0.0.8-py3-none-any.whl.

File metadata

  • Download URL: pyterminate-0.0.8-py3-none-any.whl
  • Upload date:
  • Size: 6.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.9.17

File hashes

Hashes for pyterminate-0.0.8-py3-none-any.whl
Algorithm Hash digest
SHA256 b1e89a746d09f365f098c700ac1cf3b6eb2fd561c91befca3035558c01d9a5f3
MD5 f5fd80a0e67670d181b007e9c9b08b16
BLAKE2b-256 ba86863b07c8f383b4716d950408256708c4737dcd7d412285cf0c05ecd85516

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page