Overview

inotify functionality is available from the Linux kernel and allows you to register one or more directories for watching, and to simply block and wait for notification events. This is obviously far more efficient than polling one or more directories to determine if anything has changed. This is available in the Linux kernel as of version 2.6 .

We’ve designed this library to act as a generator. All you have to do is loop, and you’ll see one event at a time and block in-between. After each cycle (all notified events were processed, or no events were received), you’ll get a None. You may use this as an opportunity to perform other tasks, if your application is being primarily driven by inotify events. By default, we’ll only block for one-second on queries to the kernel. This may be set to something else by passing a seconds-value into the constructor as block_duration_s.

Installing

Install via pip:

$ sudo pip install inotify

Example

Code:

import logging

import inotify.adapters

_DEFAULT_LOG_FORMAT = '%(asctime)s - %(name)s - %(levelname)s - %(message)s'

_LOGGER = logging.getLogger(__name__)

def _configure_logging():
    _LOGGER.setLevel(logging.DEBUG)

    ch = logging.StreamHandler()

    formatter = logging.Formatter(_DEFAULT_LOG_FORMAT)
    ch.setFormatter(formatter)

    _LOGGER.addHandler(ch)

def _main():
    i = inotify.adapters.Inotify()

    i.add_watch('/tmp')

    try:
        for event in i.event_gen():
            if event is not None:
                (header, type_names, watch_path, filename) = event
                _LOGGER.info("WD=(%d) MASK=(%d) COOKIE=(%d) LEN=(%d) MASK->NAMES=%s "
                             "WATCH-PATH=[%s] FILENAME=[%s]",
                             header.wd, header.mask, header.cookie, header.len, type_names,
                             watch_path, filename)
    finally:
        i.remove_watch('/tmp')

if __name__ == '__main__':
    _configure_logging()
    _main()

You may also choose to pass the list of directories to watch via the paths parameter of the constructor. This would work best in situations where your list of paths is static. Also, the remove_watch() call is included in the example, but is not strictly necessary. The inotify resources is cleaned-up, which would clean-up any inotify-internal watch resources as well.

Directory operations to raise events:

$ touch /tmp/aa
$ rm /tmp/aa
$ mkdir /tmp/dir1
$ rmdir /tmp/dir1

Screen output from the code, above:

2015-04-24 05:02:06,667 - __main__ - INFO - WD=(1) MASK=(256) COOKIE=(0) LEN=(16) MASK->NAMES=['IN_CREATE'] FILENAME=[aa]
2015-04-24 05:02:06,667 - __main__ - INFO - WD=(1) MASK=(32) COOKIE=(0) LEN=(16) MASK->NAMES=['IN_OPEN'] FILENAME=[aa]
2015-04-24 05:02:06,667 - __main__ - INFO - WD=(1) MASK=(4) COOKIE=(0) LEN=(16) MASK->NAMES=['IN_ATTRIB'] FILENAME=[aa]
2015-04-24 05:02:06,667 - __main__ - INFO - WD=(1) MASK=(8) COOKIE=(0) LEN=(16) MASK->NAMES=['IN_CLOSE_WRITE'] FILENAME=[aa]
2015-04-24 05:02:17,412 - __main__ - INFO - WD=(1) MASK=(512) COOKIE=(0) LEN=(16) MASK->NAMES=['IN_DELETE'] FILENAME=[aa]
2015-04-24 05:02:22,884 - __main__ - INFO - WD=(1) MASK=(1073742080) COOKIE=(0) LEN=(16) MASK->NAMES=['IN_ISDIR', 'IN_CREATE'] FILENAME=[dir1]
2015-04-24 05:02:25,948 - __main__ - INFO - WD=(1) MASK=(1073742336) COOKIE=(0) LEN=(16) MASK->NAMES=['IN_ISDIR', 'IN_DELETE'] FILENAME=[dir1]

Recursive Watching

We also provide you with the ability to add a recursive watch on a path. It turns out that there’s no low-cost way of doing this; That’s the reason that this functionality isn’t provided by the kernel. However, we recognize that this is, nonetheless, sometimes necessary.

Example:

i = inotify.adapters.InotifyTree('/tmp/watch_tree')

for event in i.event_gen():
    # Do stuff...

    pass

The only substantial difference is the type of object that was instantiated. Everything else is the same.

This will immediately recurse through the directory tree and add watches on all subdirectories. New directories will automatically have watches added for them and deleted directories will be cleaned-up.

The other differences from the standard functionality:

• You can’t remove a watch since watches are automatically managed.

• Even if you provide a very restrictive mask that doesn’t allow for directory create/delete events, the IN_ISDIR, IN_CREATE, and IN_DELETE flags will still be added.

Notes

• epoll is used to audit for inotify kernel events. This is the fastest file-descriptor “selecting” strategy.

• The design of inotify kernel-functionality is such that recursive monitoring is left as an exercise for the developer, and is not as trivial as the standard use-case.

• Due to the GIL locking considerations of Python (or any VM-based language), it is strongly recommended that, if you need to be performing other tasks while you’re concurrently watching directories, you use multiprocessing to put the directory-watching in a process of it’s own and feed information back [via queue/pipe/etc..]. This is especially true whenever your application is blocking on kernel functionality. Python’s VM will remain locked and all other threads in your application will cease to function until something raises an event in the directories that are beign watched.