Engines like Pexels provide images that one would more likely expect in a
category named ``stock_images``.
By classifying engines like Pexels more precisely, we enable a more specific
handling of them, which also helps to avoid mixing them with the more general
image search on the internet. [1]
We leave these engines *enabled* by default, take them out of ``images`` and
move them into the specialization ``stock_images``.
This makes it possible, on the one hand, for the admin to configure a tab in the
UI, and on the other hand, the user can always select the group directly using
the search syntax ``!stock_images ...``.
What we need to keep in mind:
1. Technically speaking, there are no subcategories in the strict
sense. Organizationally, the *subcategories* are derived from the
`categories_as_tabs`.
2. From an admin’s perspective, `categories_as_tabs` is the tool for
structuring their content.
3. In the context of “enabled/disabled by default”: We should avoid having
the admin have to activate many individual engines when they want to
structure their content (UI tabs).
So if we leave an engine enabled by default and **move** it from `images` to a
new category, the admin only needs to add the new category to
`categories_as_tabs` to restructure their content. If the new category is also
part of the translations (see `searx/searxng.msg` in this PR), this UI structure
would also be available internationally.
[1] https://github.com/searxng/searxng/pull/6690#issuecomment-5646291679
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
We previously set the 'cookie' cookie value with a timestamp
from 2024 to `yq=...`. The cookie is not actually used in Yandex,
instead we probably wanted to set the `yq` cookie instead. But
since the engine works without it, it can be dropped.
Also simplifies the request method a bit to be easier to read.
Code style somewhat unified, type annotations added, ``LegacyResult`` replaced
by ``Image``, but no functional modifications.
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
Queries like `!mvw hercule poirot` usually yield no results if the `description`
is not also included in the search.
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
Prevents the image proxy from decompressing brotli/gzip in transit.
See: https://curl.se/libcurl/c/CURLOPT_ACCEPT_ENCODING.html
Setting accept_encoding to none prevents curl from automatically decompressing the received contents.
Signed-off-by: vojkovic <git@vojk.au>
Additional fields:
add resolution, thumbnail and author to the image result
Types:
use EngineResults and add type hints from internall APIs
Disabled engine:
as of 09/2026: The availability of https://openverse.org/ is poor, and queries
and image requests often time out. However, these issues also occur when using
a WEB browser to search on openverse.org or view images.
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
Switching to internal APIs such as ``EngineResults`` and
adding more type hints. No functional changes.
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
The ``mkt`` argument does no longer exists and the ``async`` argument has been
renamed to ``mmasync``.
For still unknown reasons (IP based?), some users had to observe that the title
was missing [1]. I myself was not able to reproduce this error, however the
evaluation of the title was additionally expanded by the attribute value of the
``<a title=".."`` element.
Related:
- [1] https://github.com/searxng/searxng/pull/6690#issuecomment-5636811572
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
* [fix] engine: pinterest - fix empty titles and complete modernization
The titles from the fields ``title`` and ``grid_title`` are mostly empty or have
short strings without meaningful content. Various fields for the title are now
being queried, which have more informative value.
As part of the bug fix, the engine was completely revised and modernized.
Related:
- https://github.com/searxng/searxng/pull/6690#issuecomment-5631113121
Co-authored-by: @vojkovic
Fixes the bing web engine, it was just using the first word of the query for the search and return random junk other times. see: vojkovic#10
Swapped to use bing's setlang and cc params. I found us, cn, ru return complete garbage 100% of the time. I reckon that if you don't have an ip address from there it will just return garbage, so those three are skipped. Also removed accept language override because it didn't change anything anymore.
- Closes: https://github.com/searxng/searxng/issues/4964
- Related: https://github.com/vojkovic/searxng/issues/10
This has only been needed because we were flagged by HTTP2 fingerprinting.
Now, since we use curl cffi, we can bypass the fingerpinting, so we can
use HTTP2 just fine without getting blocked.
