pynzbget

pynzbget takes care of the boring parts of writing an NZBGet script, so you can focus on what your script actually does. The same script can also run under SABnzbd. It works with every NZBGet version back to v11.

Some scripts that use it: Notify, TidyIt, DirWatch and Subliminal.

Description

This provides a python framework to design NZBGet scripts with. The intent was to greatly simplify the development and debugging process. It was initially designed to work with NZBGet and is backwards compatible with all previous versions (as far back to v11). However it is also compatible with SABnzbd users too!

• It contains a built in meta tag parser to extract content from NZB-Files. Note: This can only happen if lxml is installed on your system.
• It can preform very basic obfuscation support on filenames that can not be interpreted.
• It creates a common SQLite database (optionally) to additionally write content passed via the set() function. This allows another script to later call get() and retrieve the data set() by another.
• It prepares logging right out of the box for you, there is no setup required
• All return codes have been simplified to None/True/False (you can still use the old ones if you want).
• It handles all of the exception handling. By this I mean, your code can throw an except and it's traceback will be captured gracefully to logging. Then the framework will look after returning a correct failure error code to NZBGet.
• It provides some very useful functions that are always being re-written inside of every other NZBGet script such as file scanning.
• It greatly simplifies the handling of environment variables and interaction to and from NZBGet.
• It can also be adapted to support SABnzbd scripting! Thus any script you write for NZBGet users, you can also tweak to work with SABnzbd users too!
• It runs from a lib/ folder bundled with your script, so your script works with the NZBGet v23+ Extension Manager (see below).

Install

pip install pynzbget

You need Python 3.8 or newer.

Documentation

For the most up to date information and API, visit the wiki at

• https://github.com/caronc/pynzbget/wiki

The entire framework was based on the information found here:

• NZBGet
• NZBGet Scripting Documentation
• NZBGet Extension Manager
• SABnzbd
• SABnzbd Scripting Documentation

Simplified Development

The following are some of the functionality that is built in for you:

• validate() - Handle environment checking, correct versioning as well as if the expected configuration variables you specified are present.

• health_check() - Checks the status of the retrieved content, currently this is only useful during Post-Processing.

• push() - Pushes a variables to the NZBGet server.

• set()/get() - Hash table get/set attributes that can be set in one script and then later retrieved from another. get() can also be used to fetch content that was previously pushed using the push() tool. You no longer need to work with environment variables. If you enable the SQLite database, set content is put here as well so that it can be retrieved by another script.

• unset() - This allows you to unset values set by set() and get() as well as ones set by push().

• nzb_set() - Similar to the set() function identified above except it is used to build an nzb meta hash table which can be later pushed to the server using push_dnzb().

• add_nzb() - Using the built in API/RPC NZBGet supports, this allows you to specify a path to an NZBFile which you want to enqueue for downloading.

• nzb_get() - Retieves NZB Meta information previously stored.

• nzb_unset() - Removes a variable previously set completely.

• get_statistics() - Using the built in API/RPC NZBGet supports, this retrieves and returns the statistics in an easy to ready dictionary (PostProcessScript only).

• get_logs() - Using the built in API/RPC NZBGet supports, this retrieves and returns the latest logs.

• get_files() - list all files in a specified directory as well as fetching their details such as filesize, modified date, etc in an easy to reference dictionary. You can provide a ton of different filters to minimize the content returned. Filters can by a regular expression, file prefixes, and/or suffixes.

• parse_nzbfile() - Parse an NZB-File and extract all of its meta information from it. lxml must be installed on your system for this to work correctly.

• parse_list() - Takes a string (or more) as well as lists of strings as input. It then cleans it up and produces an easy to manage list by combining all of the results into 1. Hence: parse_list('.mkv, .avi') returns: [ '.mkv', '.avi' ]

• parse_path_list() - Very smilar to parse_list() except that it is used to handle directory paths while cleaning them up at the same time.

• parse_bool() - Handles all of NZBGet's configuration options such as 'on' and 'off' as well as 'yes' or 'no', or 'True' and 'False'. It greatly simplifies the checking of these variables passed in from NZBGet.

• push_guess() - You can push a guessit dictionary (or one of your own that can help identify your release for other scripts to use later after yours finishes.

• pull_guess() - Pull a previous guess pushed by another script. why redo grunt work if it's already done for you? if no previous guess content was pushed, then an empty dictionary is returned.

• push_dnzb() - You can push all nzb meta information onbtained to the NZBGet server as DNZB_ meta tags.

• pull_dnzb() - Pull all DNZB_ meta tags issued by the server and return their values in a dictionary. if no DNZB_ (NZB Meta information) was found, then an empty dictionary is returned instead.

• deobfuscate() - Take a filename and return it in a deobfuscated to the best of its ability. (PostProcessScript only)

• is_unique_instance() - Allows you to ensure your instance of your script is unique. This is useful for Scheduled scripts which can be called and then run concurrently with NZBGet.

How To Use

• Developers are only required to define a class that inherits the NZBGet class that identifies what they are attempting to write (ScanScript, PostProcessScript, etc.).

• Then you write all of your code a the main() you must define.

Your First Script

Every NZBGet script starts with a header. NZBGet reads it to learn what kind of script you wrote and which options to show in its settings page. After that, you write a class with a main() and pynzbget does the rest.

Save this as MyExtension.py:

#!/usr/bin/env python3
###########################################################################
### NZBGET POST-PROCESSING SCRIPT

# Says hello when a download finishes.
#
# Author: Your Name.
# Version: 1.0.0.
# License: GPLv3.
# Home: https://github.com/you/MyExtension

###########################################################################
### OPTIONS

# Who to greet.
#Name=World

# Enable debug logging (yes, no).
#Debug=no

### NZBGET POST-PROCESSING SCRIPT
###########################################################################
import sys
from os.path import abspath, dirname, join

# Look in lib/ for packages first (see "NZBGet Extension Manager" below)
sys.path.insert(0, join(dirname(abspath(__file__)), 'lib'))

from nzbget import PostProcessScript


class MyExtension(PostProcessScript):
    def main(self, *args, **kwargs):
        # Make sure we're running under NZBGet and our options are set
        if not self.validate(keys=('Name', )):
            # validate() already logged why
            return False

        # Read our option (NZBGet passes it in as NZBPO_NAME)
        name = self.get('Name', 'World')

        # Read details about the download (NZBPP_NZBNAME, NZBPP_DIRECTORY)
        nzbname = self.get('NZBNAME')
        directory = self.get('DIRECTORY')

        # Logging is ready to go
        self.logger.info(f'Hello {name}! {nzbname} is in {directory}')

        # All done
        return True


if __name__ == '__main__':
    # run() calls main() and turns its answer into an NZBGet exit code
    sys.exit(MyExtension().run())

Drop it into NZBGet's scripts folder and it shows up in the settings.

What to Return

You can still use NZBGet's own exit codes if you'd rather:

from nzbget import EXIT_CODE

# Same as returning True
return EXIT_CODE.SUCCESS

Script Types

Pick the class that matches the line in your header:

Each type gets its own set of values from NZBGet. You read them all with get(), leaving off the prefix NZBGet puts in front of them.

Post Process Script Example

#############################################################################
### NZBGET POST-PROCESSING SCRIPT                                         ###
#
# Author: Your Name Goes Here <your@email.address>
#
# Describe your Post-Process Script here
#

############################################################################
### OPTIONS                                                              ###

#
# Enable NZBGet debug logging (yes, no)
#Debug=no
#

### NZBGET POST-PROCESSING SCRIPT                                         ###
#############################################################################

from nzbget import PostProcessScript

# Now define your class while inheriting the rest
class MyPostProcessScript(PostProcessScript):
    def main(self, *args, **kwargs):
        # write all of your code here you would have otherwise put in the
        # script

        if not self.validate():
            # No need to document a failure, validate will do that
            # on the reason it failed anyway
            return False

        # All system environment variables (NZBOP_.*) as well as Post
        # Process script specific content (NZBPP_.*)
        # following dictionary (without the NZBOP_ or NZBPP_ prefix):
        print('DIRECTORY %s' % self.get('DIRECTORY'))
        print('NZBNAME %s' % self.get('NZBNAME'))
        print('NZBFILENAME %s' % self.get('NZBFILENAME'))
        print('CATEGORY %s' % self.get('CATEGORY'))
        print('TOTALSTATUS %s' % self.get('TOTALSTATUS'))
        print('STATUS %s' % self.get('STATUS'))
        print('SCRIPTSTATUS %s' % self.get('SCRIPTSTATUS'))

        # Set any variable you want by any key. Note that if you use
        # keys that were defined by the system (such as CATEGORY, DIRECTORY,
        # etc, you may have some undesirable results. Try to avoid reusing
        # system variables already defined (identified above):
        self.set('MY_KEY', 'MY_VALUE')

        # You can fetch it back; this will also set an entry in  the
        # sqlite database for each hash references that can be pulled from
        # another script that simply calls self.get('MY_VAR')
        print(self.get('MY_KEY'))  # prints MY_VALUE

        # You can also use push() which is similar to set()
        # except that it interacts with the NZBGet Server and does not use
        # the sqlite database. This can only be reached across other
        # scripts if the calling application is NZBGet itself
        self.push('ANOTHER_KEY', 'ANOTHER_VALUE')

        # You can still however locally retrieve what you set using push()
        # with the get() function
        print(self.get('ANOTHER_KEY'))  # prints ANOTHER_VALUE

        # Your script configuration files (NZBPP_.*) are here in this
        # dictionary (again without the NZBPP_ prefix):
        # assume you defined `Debug=no` in the first 10K of your PostProcessScript
        # NZBGet translates this to `NZBPP_DEBUG` which can be retrieved
        # as follows:
        print('DEBUG %s' % self.get('DEBUG'))

        # Returns have been made easy. Just return:
        #   * True if everything was successful
        #   * False if there was a problem
        #   * None if you want to report that you've just gracefully
        #     skipped processing (this is better then False)
        #     in some circumstances. This is neither a failure or a
        #     success status.

        # Feel free to use the actual exit codes as well defined by
        # NZBGet on their website. They have also been defined here
        # from nzbget import EXIT_CODE

        return True

# Call your script as follows:
if __name__ == "__main__":
    from sys import exit

    # Create an instance of your Script
    ppscript = MyPostProcessScript()

    # call run() and exit() using it's returned value
    exit(ppscript.run())

Scan Script Example

############################################################################
### NZBGET SCAN SCRIPT                                                   ###
#
# Author: Your Name Goes Here <your@email.address>
#
# Describe your Scan Script here
#

############################################################################
### OPTIONS                                                              ###

#
# Enable NZBGet debug logging (yes, no)
#Debug=no
#

### NZBGET SCAN SCRIPT                                                   ###
############################################################################

from nzbget import ScanScript

# Now define your class while inheriting the rest
class MyScanScript(ScanScript):
    def main(self, *args, **kwargs):
        # write all of your code here you would have otherwise put in the
        # script

        if not self.validate():
            # No need to document a failure, validate will do that
            # on the reason it failed anyway
            return False

        # All system environment variables (NZBOP_.*) as well as Post
        # Process script specific content (NZBNP_.*)
        # following dictionary (without the NZBOP_ or NZBNP_ prefix):
        print('DIRECTORY %s' % self.get('DIRECTORY'))
        print('FILENAME %s' % self.get('FILENAME'))
        print('NZBNAME %s' % self.get('NZBNAME'))
        print('CATEGORY %s' % self.get('CATEGORY'))
        print('PRIORITY %s' % self.get('PRIORITY'))
        print('TOP %s' % self.get('TOP'))
        print('PAUSED %s' % self.get('PAUSED'))

        return True

# Call your script as follows:
if __name__ == "__main__":
    from sys import exit

    # Create an instance of your Script
    scanscript = MyScanScript()

    # call run() and exit() using it's returned value
    exit(scanscript.run())

Scheduler Script Example

############################################################################
### NZBGET SCHEDULER SCRIPT                                               ###
#
# Describe your Schedule Script here
# Author: Your Name Goes Here <your@email.address>
#

############################################################################
### OPTIONS                                                              ###

#
# Enable NZBGet debug logging (yes, no)
#Debug=no
#

### NZBGET SCHEDULER SCRIPT                                              ###
############################################################################

from nzbget import SchedulerScript

# Now define your class while inheriting the rest
class MySchedulerScript(SchedulerScript):
    def main(self, *args, **kwargs):

        # Version Checking, Environment Variables Present, etc
        if not self.validate():
            # No need to document a failure, validate will do that
            # on the reason it failed anyway
            return False

        # write all of your code here you would have otherwise put in the
        # script

        # All system environment variables (NZBOP_.*) as well as Post
        # Process script specific content (NZBSP_.*)
        # following dictionary (without the NZBOP_ or NZBSP_ prefix):
        print('DESTDIR %s' % self.get('DESTDIR'))

        return True
# Call your script as follows:
if __name__ == "__main__":
    from sys import exit

    # Create an instance of your Script
    myscript = MySchedulerScript()

    # call run() and exit() using it's returned value
    exit(myscript.run())

MultiScript Example

############################################################################
### NZBGET POST-PROCESSING/SCHEDULER SCRIPT                              ###
#
# Describe your Multi Script here
#
# Author: Your Name Goes Here <your@email.address>
#

############################################################################
### OPTIONS                                                              ###

#
# Enable NZBGet debug logging (yes, no)
#Debug=no
#

### NZBGET POST-PROCESSING/SCHEDULER SCRIPT                              ###
############################################################################

from nzbget import PostProcessScript
from nzbget import SchedulerScript

# Now define your class while inheriting the rest
class MyMultiScript(PostProcessScript, SchedulerScript):

    def postprocess_main(self, *args, **kwargs):

        # Version Checking, Environment Variables Present, etc
        if not self.validate():
            # No need to document a failure, validate will do that
            # on the reason it failed anyway
            return False

        # write your main function for your Post Processing

        return True

    def scheduler_main(self, *args, **kwargs):

        # Version Checking, Environment Variables Present, etc
        if not self.validate():
            # No need to document a failure, validate will do that
            # on the reason it failed anyway
            return False

        # write your main function for your Post Processing

        return True

# Call your script as follows:
if __name__ == "__main__":
    from sys import exit

    # Create an instance of your Script
    myscript = MyMultiScript()

    # call run() and exit() using it's returned value
    exit(myscript.run())

Queue Script

Runs when something happens in the download queue. List the events you care about under your header line:

### NZBGET QUEUE SCRIPT
### QUEUE EVENTS: NZB_ADDED, NZB_DOWNLOADED

from nzbget import QueueEvent
from nzbget import QueueScript


class MyQueueScript(QueueScript):
    def main(self, *args, **kwargs):
        if not self.validate():
            return False

        # Which event woke us up
        if self.event != QueueEvent.NZB_ADDED:
            # Not one we care about
            return None

        # The download that was just added (NZBNA_NZBNAME)
        self.logger.info(f'Queued: {self.get("NZBNAME")}')
        return True

Functions You Can Use

All of these are available as self.<name>() inside your script.

Checking Things

# Make sure we're running under NZBGet (v11 or newer) and these options
# were set in the settings page
if not self.validate(keys=('VideoPaths', 'Mode')):
    return False

# Post-processing only: did the download, repair and unpack go well?
if not self.health_check():
    return None

# Only allow one copy of this script to run at a time
if not self.is_unique_instance():
    return None

Reading and Saving Values

# Read an option, with a default if it's missing
mode = self.get('Mode', 'Preview')

# Save a value; scripts that run after yours (for the same download) can
# get() it too
self.set('MY_KEY', 'some value')

# Remove it again
self.unset('MY_KEY')

# Loop over everything get() can see
for key, value in self.items():
    self.logger.debug(f'{key} = {value}')

# Send a value to NZBGet itself so other scripts see it there too
# (NZBGet only; set() works everywhere)
self.push('ANOTHER_KEY', 'another value')

Post-processing, queue and scan scripts also keep set() values in a small SQLite database (one per download), so they're still there for the next script. Other script types can turn this on by passing database_key='my-key' when they create the script.

NZB Details

NZB-Files can carry extra details (meta tags) such as the show name or category. These work like get() and set(), but for those details:

# Read, change or remove a detail
category = self.nzb_get('category')
self.nzb_set('propername', 'My Show')
self.nzb_unset('episodename')

# Loop over them all
for key, value in self.nzb_items():
    self.logger.debug(f'{key} = {value}')

# Send our details to NZBGet (as DNZB_*), or read the ones it has
self.push_dnzb()
details = self.pull_dnzb()

# Read the details straight from an NZB-File
details = self.parse_nzbfile('/path/to/file.nzb')

push_guess() and pull_guess() do the same for "what is this download" details (such as a guessit result), so the next script doesn't have to work it out again:

# One script figures out what the download is (NZBGet only)...
self.push_guess({'type': 'episode', 'title': 'My Show', 'season': 1})

# ...and a later one reuses it (an empty dict if nothing was pushed)
guess = self.pull_guess()

Working With Files

# Find files; you get back a dict keyed by each file's path
videos = self.get_files(
    '/path/to/search',
    suffix_filter=['.mkv', '.avi'],  # only these endings
    fullstats=True,                  # include size and dates
    max_depth=2,                     # don't go too deep
)

for path, info in videos.items():
    self.logger.info(f'{path} is {info["filesize"]} bytes')

# Post-processing only: give a scrambled file name a proper one
filename = self.deobfuscate('/path/to/download/a8b7c6d5e4.mkv')

get_files() can also filter with regex_filter= and prefix_filter=. Each filter takes one value or a list.

Turning Option Text Into Python

# 'yes', 'on', 'true' and '1' become True; 'no', 'off' and so on are False
keep = self.parse_bool(self.get('KeepDirectories', 'no'))

# '.mkv, .avi .mp4' becomes a list of '.mkv', '.avi' and '.mp4'
# (duplicates are removed, so the order isn't kept)
extensions = self.parse_list(self.get('VideoExtensions'))

# Like parse_list(), but for folders (spaces in names are fine)
paths = self.parse_path_list(self.get('VideoPaths'))

# '*.iso, *.avi' becomes a list of compiled regular expressions
patterns = self.parse_regex('*.iso, *.avi', simple=True)

# Break a URL into its parts: schema, host, port, user, password, path...
url = self.parse_url('http://user:pass@localhost:8080/path')

Two more helpers live in nzbget.Utils:

from nzbget.Utils import os_path_split
from nzbget.Utils import tidy_path

# Clean up a path: tidy_path('/path//to/dir/') gives '/path/to/dir'
path = tidy_path('/path//to/dir/')

# Split a path into its parts: ['', 'path', 'to', 'dir']
parts = os_path_split('/path/to/dir')

Talking to NZBGet

These use NZBGet's API. pynzbget connects for you using NZBGet's own settings.

# Hand an NZB-File to NZBGet to download
self.add_nzb('/path/to/file.nzb', category='tv')

# Read the latest log lines
for line in self.get_logs(max_lines=100):
    self.logger.debug(line)

# Post-processing only: download, repair and unpack times and speeds
stats = self.get_statistics()
self.logger.info(f'Downloaded {stats["download_size_mb"]} MB')

SABnzbd

Write your class with SABPostProcessScript (or alongside PostProcessScript) and the same get() calls keep working. See the SABnzbd wiki page for details.

NZBGet Extension Manager

NZBGet v23 and newer can install scripts from its Extension Manager. An extension is just a folder (shipped as a zip) like this:

MyExtension/
    MyExtension.py
    manifest.json   # tells NZBGet about your script and its options
    lib/            # every package your script needs, pynzbget included

NZBGet won't run pip for you, so lib/ has to come with the script:

# Install your packages into lib/, picking versions that work on Python 3.8
pip install --target MyExtension/lib --python-version 3.8 \
    --only-binary :all: -r requirements.txt

# pynzbget works without lxml, and lxml only runs where it was built
rm -rf MyExtension/lib/lxml MyExtension/lib/lxml-*.dist-info

The sys.path.insert(...) line in the example above is what lets your script find lib/. Keep it above your other imports.

Write manifest.json once by hand, using your script's header as the guide. Keep the header too; older NZBGet versions still read it. The manifest format is documented by the NZBGet team.

Development

# Run the tests
tox -e qa

# Run the tests without lxml (how extensions run it)
tox -e minimal

# Check for lint problems (tox -e format fixes what it can)
tox -e lint

Packaging:

# Remove old builds and caches
tox -e clean

# Build the source package and wheel into dist/
tox -e build

# Build the RPMs for RHEL/Rocky 9 and 10 and Fedora 44 (needs docker);
# they land in dist/rpm/
tox -e rpm

# Do everything a release needs: clean, lint, test, build, check the
# packages and build the source RPM
tox -e release