* display the median time instead of the average.
* add a "Reliability" column (sum up the metrics and the checker results).
* the "selected language", "SafeSearch", "Time range" values are displayed as "broken" when the checker tests fail.
Report to the user suspended engines.
searx.search.processor.abstract:
* manages suspend time (per network).
* reports suspended time to the ResultContainer (method extend_container_if_suspended)
* adds the results to the ResultContainer (method extend_container)
* handles exceptions (method handle_exception)
settings.yml:
* outgoing.networks:
* can contains network definition
* propertiers: enable_http, verify, http2, max_connections, max_keepalive_connections,
keepalive_expiry, local_addresses, support_ipv4, support_ipv6, proxies, max_redirects, retries
* retries: 0 by default, number of times searx retries to send the HTTP request (using different IP & proxy each time)
* local_addresses can be "192.168.0.1/24" (it supports IPv6)
* support_ipv4 & support_ipv6: both True by default
see https://github.com/searx/searx/pull/1034
* each engine can define a "network" section:
* either a full network description
* either reference an existing network
* all HTTP requests of engine use the same HTTP configuration (it was not the case before, see proxy configuration in master)
The `url_for` function in the template context is not the one from Flask, it is
the one from `webapp`. The `webapp.url_for_theme` is different from its
namesake of Flask and has it quirks, when called with argument `_external=True`.
The `webapp.url_for_theme` can't handle absolute URLs since it pokes a leading
'/', here is the snippet of the old code::
url = url_for(endpoint, **values)
if settings['server']['base_url']:
if url.startswith('/'):
url = url[1:]
url = urljoin(settings['server']['base_url'], url)
Next drawback of (Flask's) `_external=True` is, that it will not return the HTTP
scheme when searx (the Flask app) listens on http and is proxied by a https
server.
To get the right scheme `HTTP_X_SCHEME` is needed by Flask (werkzeug). Since
this is not provided in every environment (e.g. behind Apache mod_wsgi or the
HTTP header is not fully set for some other reasons) it is recommended to
get *script_name*, *server* and *scheme* from the configured `base_url`. If
`base_url` is specified, then these values from are given preference over any
Flask's generics.
BTW this patch normalize to use `url_for` in the `opensearch.xml` and drop the
need of `host` and `urljoin` in template's context.
Signed-off-by: Markus Heiser <markus@darmarit.de>
Some of our interface locales include uppercase country codes,
which are separated by `_` instead of the more common `-`.
Also, a browser's `Accept-Language` header could be in lowercase.
This commit attempts to normalize those cases so a browser's
language+country codes can better match with our locales.
This solution assumes that our UI locales have nothing more than
language and optionally country. If we ever add a script specific
locale like `zh-Hant-TW` this would have to change to accomodate
that, but the idea would be pretty much the same as this fix.
Without this commit the module searx checks the secret_key value.
With this commit, make docs, utils/standalone_searx.py,
utils/fetch_firefox_version.py works without SEARX_DEBUG=1
For reference see https://github.com/searx/searx/pull/2386
see searx.search.processors.abstract.EngineProcessor
First the method searx call the get_params method.
If the return value is not None, then the searx call the method search.
use
from searx.engines.duckduckgo import _fetch_supported_languages, supported_languages_url # NOQA
so it is possible to easily remove all unused import using autoflake:
autoflake --in-place --recursive --remove-all-unused-imports searx tests
* URL / : the index page displayed the selected or the default category.
* URL / : when the q parameter is set using the URL, the redirect includes the URL query.
* URL /search : an empty query doesn't raise an exception.
This makes it easier to separately handle search and index requests
from a web server or from a reverse proxy.
If a request to index contains a query, a permanent redirect HTTP response
is returned. This should give some level of backwards compatibility
for users that have set a searx instance in their browser's search bar.
Xpath engine and results template changed to account for the fact that
archive.org doesn't cache .onions, though some onion engines migth have
their own cache.
Disabled by default. Can be enabled by setting the SOCKS proxies to
wherever Tor is listening and setting using_tor_proxy as True.
Requires Tor and updating packages.
To avoid manually adding the timeout on each engine, you can set
extra_proxy_timeout to account for Tor's (or whatever proxy used) extra
time.
When the user add searx as a search engine, the browser loads the /opensearch.xml URL without the cookies.
Without the query parameters, the user preferences are ignored (method and autocomplete).
In addition, opensearch.xml is modified to support automatic updates,
see https://developer.mozilla.org/en-US/docs/Web/OpenSearch
Always call initialize engines except on the first run of werkzeug with the reload feature.
the reload feature is activated when:
* searx_debug is True (SEARX_DEBUG environment variable or settings.yml)
* FLASK_APP=searx/webapp.py FLASK_ENV=development flask run (see https://flask.palletsprojects.com/en/1.1.x/cli/ )
Fix SEARX_DEBUG=0 make docs
docs/admin/engines.rst : engines are initialized
See https://github.com/searx/searx/issues/2204#issuecomment-701373438
requests 2.24.0 uses the ssl module except if it doesn't support SNI, in this case searx fallbacks to pyopenssl.
searx logs a critical message and exit if the ssl modules doesn't support SNI and pyOpenSSL is not installed.
searx logs a critical message and exit if the ssl version is older than 1.0.2.
in requirements.txt, pyopenssl is still required to install searx as a fallback.
* Log each call to get_locale: display the URL, the locale and the source (browser, preferences, form).
* Rename _get_browser_language to _get_browser_or_settings_language to match the actual code.
Sending query params over GET seems to be the only way to be able to
enable autocomplete in the browser. This commit adds the necessary URL
formatting to opensearch.xml. In order to identify queries coming from
the URL bar (rather than an AJAX request), which requires a different
JSON format and MIME type, the request headers are checked for
"X-Requested-With: XMLHttpRequest" which is added by jQuery request.
* Made first attempt at the bangs redirects plugin.
* It redirects. But in a messy way via javascript.
* First version with custom plugin
* Added a help page and a operator to see all the bangs available.
* Changed to .format because of support
* Changed to .format because of support
* Removed : in params
* Fixed path to json file and changed bang operator
* Changed bang operator back to &
* Made first attempt at the bangs redirects plugin.
* It redirects. But in a messy way via javascript.
* First version with custom plugin
* Added a help page and a operator to see all the bangs available.
* Changed to .format because of support
* Changed to .format because of support
* Removed : in params
* Fixed path to json file and changed bang operator
* Changed bang operator back to &
* Refactored getting search query. Also changed bang operator to ! and is now working.
* Removed prints
* Removed temporary bangs_redirect.js file. Updated plugin documentation
* Added unit test for the bangs plugin
* Fixed a unit test and added 2 more for bangs plugin
* Changed back to default settings.yml
* Added myself to AUTHORS.rst
* Refacored working of custom plugin.
* Refactored _get_bangs_data from list to dict to improve search speed.
* Decoupled bangs plugin from webserver with redirect_url
* Refactored bangs unit tests
* Fixed unit test bangs. Removed dubbel parsing in bangs.py
* Removed a dumb print statement
* Refactored bangs plugin to core engine.
* Removed bangs plugin.
* Refactored external bangs unit tests from plugin to core.
* Removed custom_results/bangs documentation from plugins.rst
* Added newline in settings.yml so the PR stays clean.
* Changed searx/plugins/__init__.py back to the old file
* Removed newline search.py
* Refactored get_external_bang_operator from utils to external_bang.py
* Removed unnecessary import form test_plugins.py
* Removed _parseExternalBang and _isExternalBang from query.py
* Removed get_external_bang_operator since it was not necessary
* Simplified external_bang.py
* Simplified external_bang.py
* Moved external_bangs unit tests to test_webapp.py. Fixed return in search with external_bang
* Refactored query parsing to unicode to support python2
* Refactored query parsing to unicode to support python2
* Refactored bangs plugin to core engine.
* Refactored search parameter to search_query in external_bang.py
Previously only image/jpeg was not proxied.
This commit don't proxify all MIME types starting with "image/".
This is a quick fix for the PR #1985 : the google_image engine can returns some data URL.
We have some variables in the build environment which are also needed in the
templating process. Theses variables are relavant if one creates a fork with
its own branding. We treat these variables under the term 'brands'.
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
The new url parameter "timeout_limit" set timeout limit defined in second.
Example "timeout_limit=1.5" means the timeout limit is 1.5 seconds.
In addition, the query can start with <[number] to set the timeout limit.
For number between 0 and 99, the unit is the second :
Example: "<30 searx" means the timeout limit is 3 seconds
For number above 100, the unit is the millisecond:
Example: "<850 searx" means the timeout is 850 milliseconds.
In addition, there is a new optional setting: outgoing.max_request_timeout.
If not set, the user timeout can't go above searx configuration (as before: the max timeout of selected engine for a query).
If the value is set, the user can set a timeout between 0 and max_request_timeout using
<[number] or timeout_limit query parameter.
Related to #1077
Updated version of PR #1413 from @isj-privacore
Server Timing specification: https://www.w3.org/TR/server-timing/
In the browser Dev Tools, focus on the main request, there are the responses per engine in the Timing tab.
Locale and search language was always defined with english value.
This patch inits the locale on `pre_request` in order to define the
default value of locale and language preferences.
Plus the `best_match` function provided by flask babel library did not
work as expected. So the function `match_language` provided
by searx is used to detect that the language from Accepted-Language
header can be used in searx project.
Add match_language function in utils to match any user given
language code with a list of engine's supported languages.
Also add language_aliases dict on each engine to translate
standard language codes into the custom codes used by the engine.
The exact order is
* first from SEARX_SETTINGS_PATH,
* if not found then from searx code base,
* if not found then from /etc/searx/settings.yml
* if not found an exception stops searx loading
The timeouts in settings.yml is about the total time (not only the HTTP request but also the prepare the request and parsing the response)
It was more or less the case before since the threaded_requests function ignores the thread after the timeout even the HTTP request is ended.
New / changed stats :
* page_load_time : record the HTTP request time
* page_load_count: the number of HTTP request
* engine_time : the execution total time of an engine
* engine_time_count : the number of "engine_time" measure
The avg response times in the preferences are the engine response time (engine_load_time / engine_load_count)
To sum up :
* Search.search() filters the engines that can't process the request
* Search.search() call search_multiple_requests function
* search_multiple_requests creates one thread per engine, each thread runs the search_one_request function
* search_one_request calls the request function, make the HTTP request, calls the response function, extends the result_container
* search_multiple_requests waits for the the thread to finish (or timeout)
Purposes :
- isolate the plugins calls
- distinction between parsing the web request and running the search (Search class). To be able to test code easily, to run searx code outside a web server, to filter the search query parameters with plugins more easily, etc...
Details :
- request.request_data contains request.form or request.args (initialize inside pre_request() function)
- Query class is renamed RawTextQuery
- SearchQuery class defines all search parameters
- get_search_query_from_webapp create a SearchQuery instance (basically the previous Search.__init__ code)
- Search class and SearchWithPlugins class takes a SearchQuery instance as class constructor parameter
- SearchWithPlugins class inherites from Search class, and run plugins
- A dedicated function search_with_plugins executes plugins to have a well define locals() (which is used by the plugins code).
- All plugins code is executed inside the try...except block (webapp.py, index function)
- advanced_search HTTP parameter value stays in webapp.py (it is only part of UI)
- multiple calls to result_container.get_ordered_results() doesn't compute the order multiple time (note : this method was call only once before)
- paging value is stored in the result_container class (compute in the extend method)
- test about engine.suspend_end_time is done during search method call (instead of __init__)
- check that the format parameter value is one of these : html, rss, json, rss (before the html value was assumed but some text formatting wasn't not done)
Preferences class was introduced in order to handle user preferences. Right now
it parses cookies and the form in preferences. Also it can retrieve settings
based on the name of the setting.
ATTENTION
Please note that engine preferences are handled differently from now on. So it
introduces incompatible changes. Every user who has saved preferences should reset and
save his/her settings again.
This change was needed, because everytime a default disabled engine was
added saved user preferences would broke. Now engine setting tracking is
fixed.
Alexandre Flament
Markus Heiser
Markus Heiser
Adam Tauber
Marc Abonce Seguin
renyhp
Noémi Ványi
a01200356
Émilien Devos
Mohamad Safadieh
Lukas van den Berk
Nicolas Gelot
lorddavidiii
Nicolas Gelot
rachmadaniHaryono
Venca24
yujunqiang
jibe-b
misnyo
David A Roberts
Eig8phei
stepshal
Dmitry Mikhirev
Adam Tauber