This document is work in progress; as is the searx-qt project.
searx-qt is a lightweight desktop application that lets you search on public searx instances listed on https://searx.space.
Technically searx-qt is a client application that does magic with the searx API and searx-stats2 it's instances.json
Actual search requests will be made to a server running searx software, there are many public available. We call such servers 'instances'.
If you are not familiar with the searx project; you can checkout their page here: https://asciimoo.github.io/searx/ or https://searx.me/
The searx-stats2 project lists public searx instances with statistics. The original instance is running at https://searx.space/. This is where searx-qt will request a list with instances when the update button is pressed.
name | version | license | URL |
---|---|---|---|
python | 3 | PSFL | https://docs.python.org/3/license.html |
requests | Apache 2 | http://docs.python-requests.org/en/master/ | |
PyQt5 | GPLv3 | https://www.riverbankcomputing.com/software/pyqt/intro |
Optional for socks proxy support:
name | version | license | URL |
---|---|---|---|
pysocks | BSD | https://github.com/Anorov/PySocks |
Install required dependencies:
# apt update # apt upgrade # apt install python3 python3-requests python3-pyqt5
Optional for socks proxy support:
# apt install python3-socks
Install required dependencies:
# pacman -Syu python python-requests python-pyqt5
Optional for socks proxy support:
# pacman -S python-pysocks
TODO
Request will fail on a invalid SSL/TLS certificate.
Leave checked if unsure.
See https://requests.readthedocs.io/en/master/user/advanced/#ssl-cert-verification for a more technical description.
Timeout in seconds for a single request.
Leave it at the default value of 10 seconds if unsure.
See https://requests.readthedocs.io/en/master/user/advanced/#timeouts for a more technical description.
Here you can set a proxy that will be used for every connection searx-qt makes. Leave the input field blank to use no proxy.
The 'Http' section can be used to proxy plain http:// requests to for example instances without a certificate. The 'Https' section can be used to proxy all https:// requests to instances (including fetching the instances list data from https://searx.space) with a certificate. So you can use a separate proxy for both protocols.
If you use a socks4 or socks5 proxy you probably want to make sure the 'Proxy DNS' checkbox is checked so DNS requests will also go through the proxy. DNS proxy is not available for a http proxy type.
A instance is a server running the searx project. Since we want to preform searches to searx instance(s) we need addresses of those instance(s); those will be fetched from https://searx.space/data/instances.json. The instances.json from search.space also contains a lot of other data about the instances it lists which we can use to filter instances based on our preferences.
The interface to manage instances is on the right. This is used to update the instances data with a press on the 'Update' button; filter the instances and browse/use the leftover instances in the instances table.
If searx-qt is used for the first time you will need to update the instances table. There is a 'Update' button between the Filter and the Table that can be used for this. searx-qt will not update this automatically!
It maybe useful to update the instances data so now and then since public instances appear, disappear and their stats change over time.
The instances table can be used to browse instances with their data that remain after all filters. The table is also used to set the current instance by left-clicking on one.
Right-clicking in the table opens a context-menu from where you can do the following:
The currently used instance should also be visible bottom right in the application it's status-bar.
When a filter is enabled and the instance it's value that is being matched is unknown then it is excluded by default!
Filter instances on network type. Only instances that match one of the checked network types remain.
Excludes instances that use a known MITM reverse proxy service like CloudFlare.
Exclude instances that don't have at least one IPv6 address.
Include only instances that have a searx version that has been added to this widget. Leave empty to allow all searx version.
Here are the URLs of the instances that have been manually blacklisted listed. There is a button right to each blacklist item to remove it from the blacklist.
You can manually blacklist a instance by right clicking on a instance in the instances table and click 'Add to blacklist'; multiple instances can be blacklisted at once.
If checked it will pick a random instance from the instances table if a search request fails one way or another and re-try the same request with the freshly picked instance. There is a maximum amount of 10 tries (10 different instances to try the same request on).
What is fail?
If checked it will automatically pick a random instance on a search request, it will also hide the 'Random search button' because it makes it obsolete.
If not checked it will do search requests on the same instance unless the request fails somehow and 'Fallback' is checked. Exception is when the 'Random search button' is used for the search request.
When pressed it will pick a random instance from the list and preform the search request.
When pressed it basically preforms a search request without 'Fallback' whenever it is enabled or not, it also doesn't reset the page number. So it can act as a reload button thus it's name, but it does more.
Note: When a search argument like the search query, instance URL, categories/engines etc. has changed by user interaction it will do the request with those changes, that isn't a real reload of the previous request.
Dev-note: Probably this behavior should change or the name/icon should change to something more fitting.
Preform a search request on the currently selected instance.
Page number is reset, 'Fallback' and 'Random Every' options are honored.
The query you like to search for.
See https://asciimoo.github.io/searx/user/search_syntax.html for what is possible.
It will do a search request on enter key pressed, same behavior as when the 'Search button' has been pressed.
Right clicking in (on the picture above) the dark area opens a context-menu where you can manage what options you want to be visible or not.
These are predefined by searx but can vary by instance. Example if 'Images' is checked it will only use engines with images enabled. If 'Files' is checked it will only use engines with files enabled.
searx-qt only passes this info to a searx instance with a search request, it will not automatically filter the instances table (yet).
Here you can toggle what search engines should be enabled. It will automatically filter out all instances from the instances table that doesn't have at least one of the checked engines enabled. The checked engines will be send with a search request to a searx instance with the enabled_engines param. You should only get results from engines that are checked.
If no engine is checked it means that it may return results of any engine in the list, unless one or more of the categories are checked.
The list with engines is created with data from the instances table, so only engines are listed that are available from the instances table.
Search period you like results from. Options are Last day, Last week, Last month or Last year.
If you want results in a specific language than you can select one here.