Skip to content

mugiseyebrows/event-loop

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

95 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

eventloop: Abstraction layer for filesystem events

Eventloop uses any of: pyuv or PySide2 or PyQt5 installed in the system to provide callback interface for filesystem events. And also timers.

Package intended to be a building block for utility scripts for recompiling or pushing files or restarting tests.

pyuv is used preferably as it's most efficient. uv is a library that powers nodejs event loop.

Installing

eventloop can be installed via pip as follows:

pip install eventloop

Author

Stanislav Doronin <[email protected]>

Usage

In simple case you can use on_file_changed decorator. It creates event loop, system watch and schedule, then connects them, then starts the loop and calls decorated function on filesystem events.

from eventloop import on_file_changed

if __name__ == "__main__":
    @on_file_changed("/path/to/dir")
    def your_handler(file_path):
        print(file_path)

Decorator accepts include and exclude args (list of globs or names to include or exclude files from watch) and timeout arg (read about timeout below) and loop arg.

If you need to watch more than one path you need to create and start EventLoop explicitly and pass it to decorators.

from eventloop import EventLoop, on_file_changed

if __name__ == "__main__":

    loop = EventLoop()

    @on_file_changed("/path/to/first/dir", loop=loop)
    def first_handler(file_path):
        print(file_path)

    @on_file_changed("/path/to/second/dir", loop=loop)
    def second_handler(file_path):
        print(file_path)

    loop.start()

For finer control over things you can use classes, first example can be rewriten as

from eventloop import EventLoop, FileSystemWatch, Schedule, base

class Executor(base.Executor):
    def execute(self, file_path):
        print(file_path)

if __name__ == "__main__":

    loop = EventLoop()
    executor = Executor()
    schedule = Schedule(executor)

    def on_change(file_path, event):
        schedule.append(file_path, timeout=1)

    watch = FileSystemWatch(loop)
    watch.start("/path/to/dir", on_change)
    loop.start()

Schedule caches (deduplicates) tasks appended within timeout interval, so for example three immediate consecutive changed events on same file end up in just one Executor.execute(task) call. on_file_changed decorator also uses Schedule to cache events.

Cli

You can use onchange script in command line to execute commands

python -m eventloop.onchange D:\dev\app -- echo FILE
onchange D:\dev\app -- echo FILE
onchange D:\dev\app -i *.cpp *.ui --cwd D:\dev\app\build -- ninja "&&" ctest
onchange . -i "*.pyx" --beep -- python setup.py build_ext --inplace
onchange http-server.py --server -- python -u http-server.py

License

Eventloop is distributed under the terms of MIT license, check LICENSE file.

About

Abstraction layer for filesystem events

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages