Python API and shell utilities to monitor file system events.

Works on 3.8+.

A simple program that uses watchdog to monitor directories specified as command-line arguments and logs events generated:

import sys
import time
import logging
from watchdog.observers import Observer
from watchdog.events import LoggingEventHandler

if __name__ == "__main__":
    logging.basicConfig(level=logging.INFO,
                        format='%(asctime)s - %(message)s',
                        datefmt='%Y-%m-%d %H:%M:%S')
    path = sys.argv[1] if len(sys.argv) > 1 else '.'
    logging.info(f'start watching directory {path!r}')
    event_handler = LoggingEventHandler()
    observer = Observer()
    observer.schedule(event_handler, path, recursive=True)
    observer.start()
    try:
        while True:
            time.sleep(1)
    finally:
        observer.stop()
        observer.join()

Watchdog comes with an optional utility script called watchmedo. Please type watchmedo --help at the shell prompt to know more about this tool.

Here is how you can log the current directory recursively for events related only to *.py and *.txt files while ignoring all directory events:

watchmedo log \
    --patterns="*.py;*.txt" \
    --ignore-directories \
    --recursive \
    --verbose \
    .

You can use the shell-command subcommand to execute shell commands in response to events:

watchmedo shell-command \
    --patterns="*.py;*.txt" \
    --recursive \
    --command='echo "${watch_src_path}"' \
    .

Please see the help information for these commands by typing:

watchmedo [command] --help

watchmedo can read tricks.yaml files and execute tricks within them in response to file system events. Tricks are actually event handlers that subclass watchdog.tricks.Trick and are written by plugin authors. Trick classes are augmented with a few additional features that regular event handlers don't need.

An example tricks.yaml file:

tricks:
- watchdog.tricks.LoggerTrick:
    patterns: ["*.py", "*.js"]
- watchmedo_webtricks.GoogleClosureTrick:
    patterns: ['*.js']
    hash_names: true
    mappings_format: json                  # json|yaml|python
    mappings_module: app/javascript_mappings
    suffix: .min.js
    compilation_level: advanced            # simple|advanced
    source_directory: app/static/js/
    destination_directory: app/public/js/
    files:
      index-page:
      - app/static/js/vendor/jquery*.js
      - app/static/js/base.js
      - app/static/js/index-page.js
      about-page:
      - app/static/js/vendor/jquery*.js
      - app/static/js/base.js
      - app/static/js/about-page/**/*.js

The directory containing the tricks.yaml file will be monitored. Each trick class is initialized with its corresponding keys in the tricks.yaml file as arguments and events are fed to an instance of this class as they arrive.

Install from PyPI using pip:

$ python -m pip install -U watchdog

# or to install the watchmedo utility:
$ python -m pip install -U "watchdog[watchmedo]"

Install from source:

$ python -m pip install -e .

# or to install the watchmedo utility:
$ python -m pip install -e ".[watchmedo]"

You can browse the latest release documentation online.

Fork the repository on GitHub and send a pull request, or file an issue ticket at the issue tracker. For general help and questions use stackoverflow with tag python-watchdog.

Create and activate your virtual environment, then:

python -m pip install pytest pytest-cov
python -m pip install -e ".[watchmedo]"
python -m pytest tests

If you are making a substantial change, add an entry to the "Unreleased" section of the changelog.

• Linux 2.6 (inotify)
• macOS (FSEvents, kqueue)
• FreeBSD/BSD (kqueue)
• Windows (ReadDirectoryChangesW with I/O completion ports; ReadDirectoryChangesW worker threads)
• OS-independent (polling the disk for directory snapshots and comparing them periodically; slow and not recommended)

Note that when using watchdog with kqueue, you need the number of file descriptors allowed to be opened by programs running on your system to be increased to more than the number of files that you will be monitoring. The easiest way to do that is to edit your ~/.profile file and add a line similar to:

ulimit -n 1024

This is an inherent problem with kqueue because it uses file descriptors to monitor files. That plus the enormous amount of bookkeeping that watchdog needs to do in order to monitor file descriptors just makes this a painful way to monitor files and directories. In essence, kqueue is not a very scalable way to monitor a deeply nested directory of files and directories with a large number of files.

Vim does not modify files unless directed to do so. It creates backup files and then swaps them in to replace the files you are editing on the disk. This means that if you use Vim to edit your files, the on-modified events for those files will not be triggered by watchdog. You may need to configure Vim appropriately to disable this feature.

When you want to watch changes in CIFS, you need to explicitly tell watchdog to use PollingObserver, that is, instead of letting watchdog decide an appropriate observer like in the example above, do:

from watchdog.observers.polling import PollingObserver as Observer

1. Python 3.8 or above.
2. XCode (only on macOS when installing from sources)
3. PyYAML (only for watchmedo)

Watchdog is licensed under the terms of the Apache License, version 2.0.

Copyright 2011 Yesudeep Mangalapilly.

Copyright 2012 Google, Inc & contributors.

Project source code is available at Github. Please report bugs and file enhancement requests at the issue tracker.

Too many people tried to do the same thing and none did what I needed Python to do:

• pnotify
• unison fsmonitor
• fsmonitor
• guard
• pyinotify
• inotify-tools
• jnotify
• treewatcher
• file.monitor
• pyfilesystem