Two colors are needed so I use a space in the cookie and a split in jinja to "encode" and "decode" them.
We should enforce that each theme if they must use a cookie, prefix its name with the name of the theme.
The color proposed here are based on bootstrap. We could use others.
- Modify engines to create/fetch an URL for the thumbnails
- Modify themes to show thumbnails instead of full images.
In Courgette, the result is not very beautiful. Should we change it ?
New "embedded" item for the results, allow to give an iframe to display the media directly in the results.
Note that the attributes src of the iframes are not set, but instead data-src is set, allowing to only load the iframe when clicked.
Deezer engine based on public API (no key).
Allow to perform a search while adding an engine (or a category) without adding it "officially" to the request.
'?' is used to add an engine without modifying anything else to the request. For example, you can perform a search in the 'general' category, and if you add '?tw' the result from Twitter will be added to the originals results.
Lots of Flake8 corrections
Maybe we should change the rule to allow lines of 120 chars. It seems more usable.
Big twitter correction : now it outputs the words in right order...
Allow the user to select a language. It must be written in english, and capitalized, ie : English, French, German, Hungarian... (reverted from commit 829948b85d)
Allow to search for images on 500px. It doesn't use the official API, but the page result. Less stable, but less API key to possess...
Two engines were necessary for Searchcode because there are to search mode : search for documentation or search for code example. Both use open APIs.
It uses the webpage json infos to build the results
Let the user choose the engine in setting.yml. Noapi active by default
+ little corrections on Flickr engine
Everything was redone to use the API. It needs an API key, but it's worth it.
Everything works. Title, Image, Content, URL
The API allow lots of things. Thumbnails and date will be easy to add when it will be implemented in Searx.
Fixasciimoo/searx#126
Everything was redone to use the API. It needs an API key, but it's worth it.
Everything works. Title, Image, Content, URL
The API allow lots of things. Thumbnails and date will be easy to add when it will be implemented in Searx.
Fixasciimoo/searx#126
Parse and return results correctly. Pages numbers taken care of.
Not done, and maybe to do :
- 'content' : I don't know what it could be. Maybe votes ?
- 'categories' : the results are not filtered by categories, because I don't see how to do it properly : there are too much categories on Kickass. Is 'video' only movies, or also tv show or porn ? So for now, the category is 'all'.
- Favicon/icon : may be a good idea.
For DoB and DoD, wikipedia use a non standard ISO format, not easily readable.
Now the date is displayed in an human readable form, using the language setting as locale if available. If not, it uses the default locale.
It seems like Google image is doing a double urlencode on the url of the images. So we need to unquote once before sending to the browser the urls.
It solves the 404 we could see with some image with specials chars in url.
Exemple https://searx.laquadrature.net/?q=etes&pageno=1&category_images (there are two of those in the list)
Cqoicebordel