The previous method of removing all site filters from the search query
removed the last letter of the search. This only applies the substring
filter if any site filters are present in the query.
Fixes#306
* Block websites in search results via user config
Adds a new config field "Block" to specify a comma separated list of
websites to block in search results. This is applied for all searches.
* Add test for blocking sites from search results
* Document WHOOGLE_CONFIG_BLOCK usage
* Strip '-site:' filters from query in header template
The 'behind the scenes' site filter applied for blocked sites was
appearing in the query field when navigating between search categories
(all -> images -> news, etc). This prevents the filter from appearing in
all except "images", since the image category uses a separate header.
This should eventually be addressed when the image page can begin using
the standard whoogle header, but until then, the filter will still
appear for image searches.
The env_file option is a better way of loading environment variables
from whoogle.env, and has been added to docker-compose.yml.
The whoogle.env file had comments after each example setting, and it was
not clear that these comments needed to be removed when setting new
values. The comments for each variable have been moved above the
appropriate variable to reduce confusion.
See #303 for initial discussion
* Add option to disable changing of configuration
Introduces a test to ensure the correct response code is found when
attempting to update the config when disabled, and ensure default config
is unchanged when posting a new config dict.
Attempting to update the config using the API when disabled now returns
a 403 code + redirect.
Co-authored-by: Ben Busby <benbusby@protonmail.com>
The search language is now set using the WHOOGLE_CONFIG_SEARCH_LANGUAGE
environment variable. Interface language is still set using
WHOOGLE_CONFIG_LANGUAGE.
Fixes#260
Enforces 0 margin for the search input form on the result page, which
removes the weird gap that is seen by default.
Also made minor changes to the border styling. Desktop searches now have
a single bottom border in dark mode rather than an all around border,
and the border around the mobile search result input was removed
entirely.
This was unfortunately a bit more complex than just adding an HTML reset
button, since reset buttons only "reset" input content to its original
value rather than clearing it. This doesn't work for Whoogle's needs,
since inputs on search result pages are auto populated with the search
content as their default value.
A reset button was introduced anyways, but is controlled by a few lines
of javascript to allow completely clearing the search input. The button
will only appear on mobile searches.
At the moment, it isn't particularly pretty, but is functional. It uses
just a plain "x" character and is always visible on mobile search result
pages. This leaves plenty of room for improvement moving forward.
Fixes#291
The recent change to cast bool config vars as ints to handle a '0' or
'1' value was shortsighted, since it doesn't allow for instances where
the variable is set to an empty value (or '' or any invalid/non-int
value).
This introduces a read_config_bool method for reading values that should
be a '0' or '1', but will default to False if not a digit (otherwise the
value will be cast as bool(int(value)) if "value" is a digit str).
Fixes#288
Config boolean environment variables need to be cast to ints, since
they are set or unset using 0 and 1. Previously they were interpreted as
(pseudocode) read_var(name, default=False), which meant that setting
CONFIG_VAR=0 would enable that variable since Python reads environment
variables as strings, and '0' is truthy. This updates the previous logic
to (still pseudocode) int(read_var(name, default='0')).
Fixes#279
Both light and dark themes have been updated to remove the leftover
hardcoded values (mostly related to the search suggestion styling).
See discussion in #247.
wget's default behavior to download contents to a file is not always a
valid method to use as a Docker healthcheck, due to permission issues.
This circumvents any issues there by redirecting the healthcheck output
to stdout, which shouldn't cause any issues regardless of context.
Fixes#272
The logging from imported modules (stem, in particular) has caused quite
a few users to assume there are errors where there aren't any. The logs
from stem also aren't helpful, as everything in the library works as
expected despite the implication from the logs that it is not working.
Randomizing the "Mozilla" portion of the user agent changed the
character encoding to GB2312. Setting it to plain "Mozilla" enforces
UTF-8 encoding.
Bump to version 0.4.1 for release of bug fix
Fixes#267
This moves away from the previous (messy) approach of using two separate
keys for decrypting text and element URLs separately and regenerating
them for new searches. The current implementation of sessions is not very
reliable, which lead to keys being regenerated too soon, which would
break page navigation. Until that can be addressed, the single
key per session approach should work a lot better.
Fixes#250Fixes#90
The previous implementation of the is_heroku check in
search.needs_https() was implemented to only match URLs ending in
'.herokuapp.com', and skipped upgrading to HTTPS for other endpoints.
