-
Notifications
You must be signed in to change notification settings - Fork 464
How to write a search plugin
qBittorrent provides a search engine plugins management system. Thanks to this, you can easily write your own plugins to look for torrents in your favorite Bittorrent search engines and extend qBittorrent integrated search engine.
- All you need is some motivation and some knowledge of Python language.
- The minimum supported python version is specified here, make sure your plugin can work with it and every later versions.
-
Only import libraries from Python Standard Library.
Third party libraries (such as those installed from PyPI) are not guaranteed to be present in user's environment. - You are encouraged to ensure good quality of your plugin: Python Code Quality: Tools & Best Practices.
For example, here is how the official plugins are checked: ci.yaml.
3.1 Code Examples
3.5 Notes
First, you must understand that a qBittorrent search engine plugin is actually a Python class file whose task is to contact a search engine website (e.g. Mininova.org), parse the results displayed by the web page and print them on stdout with the following syntax:
link|name|size|seeds|leech|engine_url|desc_link
One search result per line.
For example:
https://www.mininova.org/get/1109509|ubuntu-hardy-desktop-i386-alpha-3 iso|711332986|0|0|https://www.mininova.org
https://www.mininova.org/get/1088604|Xubuntu 7.10 for Virtual PC 2007|713901998|0|0|https://www.mininova.org
Your plugin should be named "engine_name.py", in lowercase and without spaces nor any special characters. You'll also need the other files for the project (Link)
The Files Are:
-> nova2.py # the main search engine script which calls the plugins
-> nova2dl.py # standalone script called by qBittorrent to download a torrent using a particular search plugin
-> helpers.py # contains helper functions you can use in your plugins such as retrieve_url() and download_file()
-> novaprinter.py # contains some useful functions like prettyPrint(my_dict) to display your search results
-> socks.py # Required by helpers.py. This module provides a standard socket-like interface.
Here is the basic structure of engine_name.py:
#VERSION: 1.00
# AUTHORS: YOUR_NAME (YOUR_MAIL)
# LICENSING INFORMATION
from html.parser import HTMLParser
from helpers import download_file, retrieve_url
from novaprinter import prettyPrinter
# some other imports if necessary
class engine_name(object):
"""
`url`, `name`, `supported_categories` should be static variables of the engine_name class,
otherwise qbt won't install the plugin.
`url`: The URL of the search engine.
`name`: The name of the search engine, spaces and special characters are allowed here.
`supported_categories`: What categories are supported by the search engine and their corresponding id,
possible categories are ('all', 'anime', 'books', 'games', 'movies', 'music', 'pictures', 'software', 'tv').
"""
url = 'https://www.engine-url.org'
name = 'Full engine name'
supported_categories = {
'all': '0',
'anime': '7',
'games': '2',
'movies': '6',
'music': '1',
'software': '3',
'tv': '4'
}
def __init__(self):
"""
Some initialization
"""
def download_torrent(self, info):
"""
Providing this function is optional.
It can however be interesting to provide your own torrent download
implementation in case the search engine in question does not allow
traditional downloads (for example, cookie-based download).
"""
print(download_file(info))
# DO NOT CHANGE the name and parameters of this function
# This function will be the one called by nova2.py
def search(self, what, cat='all'):
"""
Here you can do what you want to get the result from the search engine website.
Everytime you parse a result line, store it in a dictionary
and call the prettyPrint(your_dict) function.
`what` is a string with the search tokens, already escaped (e.g. "Ubuntu+Linux")
`cat` is the name of a search category in ('all', 'anime', 'books', 'games', 'movies', 'music', 'pictures', 'software', 'tv')
"""PLEASE note that the filename (without .py extension) must be identical to the class name. Otherwise, qBittorrent will refuse to install it!
After downloading the content of the web page containing the results (using retrieve_url()), you will want to parse it in order to create a dict per search result and call prettyPrint(your_dict) function to display it on stdout (in a format understandable by qBittorrent).
In order to parse the pages, you can use the following python modules (not exhaustive):
- [ADVISED METHOD] HTMLParser / html.parser: Builtin python parser which replaces deprecated SGMLParser. Mostly similar to the SMGLParser
-
xml.dom.minidom: XML parser. Be careful, this parser is very sensitive and the website must be fully XHTML compliant for this to work. -
re: If you like using regular expressions (regex)
Note that the size is in provided bytes.
To achieve this task we provide several helper functions such as prettyPrinter().
In fact, you don't really need to pay attention to the output syntax because we provide a function for this called prettyPrinter(dictionary). You can import it using the following command:
from novaprinter import prettyPrinterYou must pass to this function a dictionary containing the following keys (value should be -1 if you do not have the info):
-
link=> A string corresponding the the download link (the .torrent file or magnet link) -
name=> A unicode string corresponding to the torrent's name (i.e: "Ubuntu Linux v6.06") -
size=> A string corresponding to the torrent size (i.e: "6 MB" or "200 KB" or "1.2 GB"...) -
seeds=> The number of seeds for this torrent (as a string) -
leech=> The number of leechers for this torrent (a a string) -
engine_url=> The search engine url (i.e: https://www.mininova.org) -
desc_link=> A string corresponding to the the description page for the torrent
The retrieve_url() method takes an URL as parameter and returns the content of the URL as a string.
This function is useful to get the search results from a Bittorrent search engine site. All you need to do is to pass the properly formatted URL to the function (the URL usually include GET parameters relative to search tokens, category, sorting, page number).
from helpers import retrieve_url
dat = retrieve_url(self.url + '/search?q=%s&c=%s&o=52&p=%d' % (what, self.supported_categories[cat], i))The download_file() functions takes as a parameter the URL to a torrent file. This function will download the torrent to a temporary location and print on stdout:
path_to_temporary_file urlIt prints two values separated by a space:
- The path to the downloaded file (usually in /tmp folder)
- The URL from which the file was downloaded
Here is an example:
from helpers import retrieve_url, download_file
print download_file(url)
> /tmp/esdzes https://www.mininova.org/get/123456Do not hesitate to use the official search engine plugins as an example. They are available here.
- kickasstorrents.py uses json module
- torrentreactor.py uses HTMLParser module
Before installing your plugin (in Qbittorrent) you can test run the plugin while debugging it. Hence, we advise that you download these files.
You will get the following structure:
your_search_engine
-> nova2.py # the main search engine script which calls the plugins
-> nova2dl.py # standalone script called by qBittorrent to download a torrent using a particular search plugin
-> helpers.py # contains helper functions you can use in your plugins such as retrieve_url() and download_file()
-> novaprinter.py # contains some useful functions like prettyPrint(my_dict) to display your search results
-> socks.py # Required by helpers.py. This module provides a standard socket-like interface.
Put your plugin in engines/ folder ( %localappdata%\qBittorrent\nova3\engines\ ) and then in CMD execute nova2.py script like this:
..\nova2.py your_search_engine_name category search_tokens
# e.g.: ..\nova2.py mininova all kubuntu linux
# e.g.: ..\nova2.py btjunkie books ubuntuA successful result will output:
DEBUG:root:C:\users\user\appdata\local\qbittorrent\nova3\qbt\qbt
the app will start listing links it finds in the following format:
link|name|size|#seeds|#leechers|engine|page url
- Go to search tab in main window, click on "Search engines..." button.
- Then, a new window will pop up, containing the list of installed search engine plugins.
- Click on "Install a new one" at the bottom and select your
*.pypython script on you filesystem.
If everything goes well, qBittorrent should notify you that it was successfully installed and your plugin should appear in the list.
Once you managed to write a search engine plugin for qBittorrent that works, feel free to post it on this wiki page so that the other users can use it too.
If you are lucky, your plugin may also be included in the official repository.
- As a convention, it is advised that you print the results sorted by number of seeds (the most seeds at the top) because these are usually the most interesting torrents.
- Please note that search engines usually display results on several pages. Hence, it is better to parse all these pages to get all results. All official plugins have multipage support.
- Some search engines do not provide all the informations required by
prettyPrinter(). If it is the case, set-1as value for the given key (i.e.:torrent_info['seeds'] = -1) - Plugins packaged in a python are no longer directly installable since qBittorrent v2.0.0. You must provide qBittorrent with the python file.