summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
-rw-r--r--Makefile4
-rw-r--r--docs/admin/installation-apache.rst15
-rw-r--r--docs/admin/installation-nginx.rst10
-rw-r--r--docs/admin/morty.rst15
-rw-r--r--requirements.txt2
-rw-r--r--searx/engines/google.py15
-rw-r--r--searx/engines/google_images.py12
-rw-r--r--searx/engines/google_news.py220
-rw-r--r--searx/search/checker/__main__.py1
-rw-r--r--searx/search/checker/impl.py15
-rw-r--r--searx/search/processors/online.py6
-rw-r--r--searx/settings.yml12
-rwxr-xr-xutils/lxc.sh1
-rwxr-xr-xutils/searx.sh6
14 files changed, 229 insertions, 105 deletions
diff --git a/Makefile b/Makefile
index 9d8e3a199..350b2321b 100644
--- a/Makefile
+++ b/Makefile
@@ -177,7 +177,9 @@ PYLINT_FILES=\
searx/testing.py \
searx/engines/gigablast.py \
searx/engines/deviantart.py \
- searx/engines/digg.py
+ searx/engines/digg.py \
+ searx/engines/google.py \
+ searx/engines/google_news.py
test.pylint: pyenvinstall
$(call cmd,pylint,$(PYLINT_FILES))
diff --git a/docs/admin/installation-apache.rst b/docs/admin/installation-apache.rst
index 217e57718..311b5c250 100644
--- a/docs/admin/installation-apache.rst
+++ b/docs/admin/installation-apache.rst
@@ -180,10 +180,6 @@ modules and create a `Location`_ configuration for the searx site. In most
distributions you have to un-comment the lines in the main configuration file,
except in :ref:`The Debian Layout`.
-To pass the HTTP HOST header
-With ProxyPreserveHost_ the incoming Host HTTP request header is passed to the
-proxied host.
-
.. tabs::
.. group-tab:: Ubuntu / debian
@@ -231,6 +227,11 @@ proxied host.
LoadModule proxy_module modules/mod_proxy.so
LoadModule proxy_http_module modules/mod_proxy_http.so
+With ProxyPreserveHost_ the incoming Host HTTP request header is passed to the
+proxied host.
+
+.. _apache searx via filtron plus morty:
+
.. tabs::
.. group-tab:: searx via filtron plus morty
@@ -285,15 +286,15 @@ proxied host.
</Location>
- Note that reverse proxy advised to be used in case of single-user or
- low-traffic instances. For a fully result proxification add :ref:`morty's
- <searx morty>` **public URL** to your :origin:`searx/settings.yml`:
+ For a fully result proxification add :ref:`morty's <searx morty>` **public
+ URL** to your :origin:`searx/settings.yml`:
.. code:: yaml
result_proxy:
# replace example.org with your server's public name
url : https://example.org/morty
+ key : !!binary "insert_your_morty_proxy_key_here"
server:
image_proxy : True
diff --git a/docs/admin/installation-nginx.rst b/docs/admin/installation-nginx.rst
index 589c40ada..430ebbcce 100644
--- a/docs/admin/installation-nginx.rst
+++ b/docs/admin/installation-nginx.rst
@@ -163,6 +163,8 @@ Started wiki`_ is always a good resource *to keep in the pocket*.
Create configuration at ``/etc/nginx/conf.d/searx`` and place a
symlink to sites-enabled:
+.. _nginx searx via filtron plus morty:
+
.. tabs::
.. group-tab:: searx via filtron plus morty
@@ -210,21 +212,21 @@ Started wiki`_ is always a good resource *to keep in the pocket*.
proxy_set_header X-Scheme $scheme;
}
- Note that reverse proxy advised to be used in case of single-user or
- low-traffic instances. For a fully result proxification add :ref:`morty's
- <searx morty>` **public URL** to your :origin:`searx/settings.yml`:
+ For a fully result proxification add :ref:`morty's <searx morty>` **public
+ URL** to your :origin:`searx/settings.yml`:
.. code:: yaml
result_proxy:
# replace example.org with your server's public name
url : https://example.org/morty
+ key : !!binary "insert_your_morty_proxy_key_here"
server:
image_proxy : True
- .. group-tab:: proxy or uWSGI
+ .. group-tab:: proxy or uWSGI
Be warned, with this setup, your instance isn't :ref:`protected <searx
filtron>`. Nevertheless it is good enough for intranet usage and it is a
diff --git a/docs/admin/morty.rst b/docs/admin/morty.rst
index 2858fde3c..5468c9c13 100644
--- a/docs/admin/morty.rst
+++ b/docs/admin/morty.rst
@@ -16,15 +16,22 @@ By default searx can only act as an image proxy for result images, but it is
possible to proxify all the result URLs with an external service, morty_.
To use this feature, morty has to be installed and activated in searx's
-``settings.yml``.
-
-Add the following snippet to your ``settings.yml`` and restart searx:
+``settings.yml``. Add the following snippet to your ``settings.yml`` and
+restart searx:
.. code:: yaml
result_proxy:
url : http://127.0.0.1:3000/
- key : your_morty_proxy_key
+ key : !!binary "insert_your_morty_proxy_key_here"
+
+Note that the example above (``http://127.0.0.1:3000``) is only for single-user
+instances without a HTTP proxy. If your morty service is public, the url is the
+address of the reverse proxy (e.g ``https://example.org/morty``).
+
+For more information about *result proxy* have a look at *"searx via filtron
+plus morty"* in the :ref:`nginx <nginx searx via filtron plus morty>` and
+:ref:`apache <apache searx via filtron plus morty>` sections.
``url``
Is the address of the running morty service.
diff --git a/requirements.txt b/requirements.txt
index 776bbc20b..a8d9b3f20 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -9,4 +9,4 @@ pygments==2.1.3
python-dateutil==2.8.1
pyyaml==5.3.1
requests[socks]==2.25.1
-pycld3==0.20
+langdetect==1.0.8
diff --git a/searx/engines/google.py b/searx/engines/google.py
index 4198de640..fe9cd63e0 100644
--- a/searx/engines/google.py
+++ b/searx/engines/google.py
@@ -155,6 +155,11 @@ def get_lang_country(params, lang_list, custom_aliases):
return language, country, lang_country
+def detect_google_sorry(resp):
+ resp_url = urlparse(resp.url)
+ if resp_url.netloc == 'sorry.google.com' or resp_url.path.startswith('/sorry'):
+ raise SearxEngineCaptchaException()
+
def request(query, params):
"""Google search request"""
@@ -200,16 +205,10 @@ def request(query, params):
def response(resp):
"""Get response from google's search request"""
- results = []
- # detect google sorry
- resp_url = urlparse(resp.url)
- if resp_url.netloc == 'sorry.google.com' or resp_url.path == '/sorry/IndexRedirect':
- raise SearxEngineCaptchaException()
-
- if resp_url.path.startswith('/sorry'):
- raise SearxEngineCaptchaException()
+ detect_google_sorry(resp)
+ results = []
# which subdomain ?
# subdomain = resp.search_params.get('google_subdomain')
diff --git a/searx/engines/google_images.py b/searx/engines/google_images.py
index 8c2cb9d2a..612682c44 100644
--- a/searx/engines/google_images.py
+++ b/searx/engines/google_images.py
@@ -12,10 +12,9 @@ Definitions`_.
Header set Content-Security-Policy "img-src 'self' data: ;"
"""
-from urllib.parse import urlencode, urlparse, unquote
+from urllib.parse import urlencode, unquote
from lxml import html
from searx import logger
-from searx.exceptions import SearxEngineCaptchaException
from searx.utils import extract_text, eval_xpath
from searx.engines.google import _fetch_supported_languages, supported_languages_url # NOQA # pylint: disable=unused-import
@@ -23,6 +22,7 @@ from searx.engines.google import (
get_lang_country,
google_domains,
time_range_dict,
+ detect_google_sorry,
)
logger = logger.getChild('google images')
@@ -123,13 +123,7 @@ def response(resp):
"""Get response from google's search request"""
results = []
- # detect google sorry
- resp_url = urlparse(resp.url)
- if resp_url.netloc == 'sorry.google.com' or resp_url.path == '/sorry/IndexRedirect':
- raise SearxEngineCaptchaException()
-
- if resp_url.path.startswith('/sorry'):
- raise SearxEngineCaptchaException()
+ detect_google_sorry(resp)
# which subdomain ?
# subdomain = resp.search_params.get('google_subdomain')
diff --git a/searx/engines/google_news.py b/searx/engines/google_news.py
index 63fef6696..e83b2ba48 100644
--- a/searx/engines/google_news.py
+++ b/searx/engines/google_news.py
@@ -1,12 +1,45 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
-"""
- Google (News)
+"""Google (News)
+
+For detailed description of the *REST-full* API see: `Query Parameter
+Definitions`_. Not all parameters can be appied, e.g. num_ (the number of
+search results to return) is ignored.
+
+.. _Query Parameter Definitions:
+ https://developers.google.com/custom-search/docs/xml_results#WebSearch_Query_Parameter_Definitions
+
+.. _num: https://developers.google.com/custom-search/docs/xml_results#numsp
+
"""
+# pylint: disable=invalid-name, missing-function-docstring
+
+import binascii
+import re
from urllib.parse import urlencode
+from base64 import b64decode
from lxml import html
-from searx.utils import match_language
-from searx.engines.google import _fetch_supported_languages, supported_languages_url # NOQA # pylint: disable=unused-import
+
+from searx import logger
+from searx.utils import (
+ eval_xpath,
+ eval_xpath_list,
+ eval_xpath_getindex,
+ extract_text,
+)
+
+# pylint: disable=unused-import
+from searx.engines.google import (
+ supported_languages_url,
+ _fetch_supported_languages,
+ detect_google_sorry,
+)
+# pylint: enable=unused-import
+
+from searx.engines.google import (
+ get_lang_country,
+ filter_mapping,
+)
# about
about = {
@@ -18,72 +51,147 @@ about = {
"results": 'HTML',
}
-# search-url
+logger = logger.getChild('google news')
+
+# compared to other google engines google-news has a different time range
+# support. The time range is included in the search term.
+time_range_dict = {
+ 'day': 'when:1d',
+ 'week': 'when:7d',
+ 'month': 'when:1m',
+ 'year': 'when:1y',
+}
+
+# engine dependent config
+
categories = ['news']
-paging = True
+paging = False
language_support = True
-safesearch = True
+use_locale_domain = True
time_range_support = True
-number_of_results = 10
-
-search_url = 'https://www.google.com/search'\
- '?{query}'\
- '&tbm=nws'\
- '&gws_rd=cr'\
- '&{search_options}'
-time_range_attr = "qdr:{range}"
-time_range_dict = {'day': 'd',
- 'week': 'w',
- 'month': 'm',
- 'year': 'y'}
-
+safesearch = True # not really, but it is not generated by google
-# do search-request
def request(query, params):
-
- search_options = {
- 'start': (params['pageno'] - 1) * number_of_results
- }
-
- if params['time_range'] in time_range_dict:
- search_options['tbs'] = time_range_attr.format(range=time_range_dict[params['time_range']])
-
- if safesearch and params['safesearch']:
- search_options['safe'] = 'on'
-
- params['url'] = search_url.format(query=urlencode({'q': query}),
- search_options=urlencode(search_options))
-
- if params['language'] != 'all':
- language = match_language(params['language'], supported_languages, language_aliases).split('-')[0]
- if language:
- params['url'] += '&hl=' + language
+ """Google-News search request"""
+
+ language, country, lang_country = get_lang_country(
+ # pylint: disable=undefined-variable
+ params, supported_languages, language_aliases
+ )
+ subdomain = 'news.google.com'
+
+ if params['time_range']: # in time_range_dict:
+ query += ' ' + time_range_dict[params['time_range']]
+
+ query_url = 'https://'+ subdomain + '/search' + "?" + urlencode({
+ 'q': query,
+ 'hl': language,
+ 'lr': "lang_" + language,
+ 'ie': "utf8",
+ 'oe': "utf8",
+ 'ceid' : "%s:%s" % (country, language),
+ 'gl' : country,
+ })
+
+ if params['safesearch']:
+ query_url += '&' + urlencode({'safe': filter_mapping[params['safesearch']]})
+
+ params['url'] = query_url
+ logger.debug("query_url --> %s", query_url)
+
+ # en-US,en;q=0.8,en;q=0.5
+ params['headers']['Accept-Language'] = (
+ lang_country + ',' + language + ';q=0.8,' + language + ';q=0.5'
+ )
+ logger.debug("HTTP header Accept-Language --> %s",
+ params['headers']['Accept-Language'])
+ params['headers']['Accept'] = (
+ 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8'
+ )
+
+ # hl=en redirect to hl=en-US / en-CA ...
+ params['soft_max_redirects'] = 1
+
+ #params['google_subdomain'] = subdomain
return params
-# get response from search-request
def response(resp):
+ """Get response from google's search request"""
results = []
+ detect_google_sorry(resp)
+
+ # which subdomain ?
+ # subdomain = resp.search_params.get('google_subdomain')
+
+ # convert the text to dom
dom = html.fromstring(resp.text)
- # parse results
- for result in dom.xpath('//div[@class="g"]|//div[@class="g _cy"]'):
- try:
- r = {
- 'url': result.xpath('.//a[@class="l lLrAF"]')[0].attrib.get("href"),
- 'title': ''.join(result.xpath('.//a[@class="l lLrAF"]//text()')),
- 'content': ''.join(result.xpath('.//div[@class="st"]//text()')),
- }
- except:
- continue
-
- imgs = result.xpath('.//img/@src')
- if len(imgs) and not imgs[0].startswith('data'):
- r['img_src'] = imgs[0]
-
- results.append(r)
+ for result in eval_xpath_list(dom, '//div[@class="xrnccd"]'):
+
+ # The first <a> tag in the <article> contains the link to the
+ # article The href attribute of the <a> is a google internal link,
+ # we can't use. The real link is hidden in the jslog attribute:
+ #
+ # <a ...
+ # jslog="95014; 4:https://www.cnn.com/.../index.html; track:click"
+ # href="./articles/CAIiENu3nGS...?hl=en-US&amp;gl=US&amp;ceid=US%3Aen"
+ # ... />
+
+ jslog = eval_xpath_getindex(result, './article/a/@jslog', 0)
+ url = re.findall('http[^;]*', jslog)
+ if url:
+ url = url[0]
+ else:
+ # The real URL is base64 encoded in the json attribute:
+ # jslog="95014; 5:W251bGwsbnVsbCxudW...giXQ==; track:click"
+ jslog = jslog.split(";")[1].split(':')[1].strip()
+ try:
+ padding = (4 -(len(jslog) % 4)) * "="
+ jslog = b64decode(jslog + padding)
+ except binascii.Error:
+ # URL cant be read, skip this result
+ continue
+
+ # now we have : b'[null, ... null,"https://www.cnn.com/.../index.html"]'
+ url = re.findall('http[^;"]*', str(jslog))[0]
+
+ # the first <h3> tag in the <article> contains the title of the link
+ title = extract_text(eval_xpath(result, './article/h3[1]'))
+
+ # the first <div> tag in the <article> contains the content of the link
+ content = extract_text(eval_xpath(result, './article/div[1]'))
+
+ # the second <div> tag contains origin publisher and the publishing date
+
+ pub_date = extract_text(eval_xpath(result, './article/div[2]//time'))
+ pub_origin = extract_text(eval_xpath(result, './article/div[2]//a'))
+
+ pub_info = []
+ if pub_origin:
+ pub_info.append(pub_origin)
+ if pub_date:
+ # The pub_date is mostly a string like 'yesertday', not a real
+ # timezone date or time. Therefore we can't use publishedDate.
+ pub_info.append(pub_date)
+ pub_info = ', '.join(pub_info)
+ if pub_info:
+ content = pub_info + ': ' + content
+
+ # The image URL is located in a preceding sibling <img> tag, e.g.:
+ # "https://lh3.googleusercontent.com/DjhQh7DMszk.....z=-p-h100-w100"
+ # These URL are long but not personalized (double checked via tor).
+
+ img_src = extract_text(result.xpath('preceding-sibling::a/figure/img/@src'))
+
+ results.append({
+ 'url': url,
+ 'title': title,
+ 'content': content,
+ 'img_src': img_src,
+ })
# return results
return results
diff --git a/searx/search/checker/__main__.py b/searx/search/checker/__main__.py
index 75b37e6c5..0d7d1b8ed 100644
--- a/searx/search/checker/__main__.py
+++ b/searx/search/checker/__main__.py
@@ -74,6 +74,7 @@ def run(engine_name_list, verbose):
stdout.write(f' {"found languages":15}: {" ".join(sorted(list(checker.test_results.languages)))}\n')
for test_name, logs in checker.test_results.logs.items():
for log in logs:
+ log = map(lambda l: l if isinstance(l, str) else repr(l), log)
stdout.write(f' {test_name:15}: {RED}{" ".join(log)}{RESET_SEQ}\n')
diff --git a/searx/search/checker/impl.py b/searx/search/checker/impl.py
index 244536f1b..25887b0f4 100644
--- a/searx/search/checker/impl.py
+++ b/searx/search/checker/impl.py
@@ -9,7 +9,8 @@ from time import time
from urllib.parse import urlparse
import re
-import cld3
+from langdetect import detect_langs
+from langdetect.lang_detect_exception import LangDetectException
import requests.exceptions
from searx import poolrequests, logger
@@ -181,10 +182,14 @@ class ResultContainerTests:
self.test_results.add_error(self.test_name, message, *args, '(' + sqstr + ')')
def _add_language(self, text: str) -> typing.Optional[str]:
- r = cld3.get_language(str(text)) # pylint: disable=E1101
- if r is not None and r.probability >= 0.98 and r.is_reliable:
- self.languages.add(r.language)
- self.test_results.add_language(r.language)
+ try:
+ r = detect_langs(str(text)) # pylint: disable=E1101
+ except LangDetectException:
+ return None
+
+ if len(r) > 0 and r[0].prob > 0.95:
+ self.languages.add(r[0].lang)
+ self.test_results.add_language(r[0].lang)
return None
def _check_result(self, result):
diff --git a/searx/search/processors/online.py b/searx/search/processors/online.py
index 0ceb0adf2..d79edd542 100644
--- a/searx/search/processors/online.py
+++ b/searx/search/processors/online.py
@@ -239,14 +239,14 @@ class OnlineProcessor(EngineProcessor):
'test': ['unique_results']
}
- if getattr(self.engine, 'lang', False):
+ if getattr(self.engine, 'supported_languages', []):
tests['lang_fr'] = {
'matrix': {'query': 'paris', 'lang': 'fr'},
- 'result_container': ['not_empty', ('has_lang', 'fr')],
+ 'result_container': ['not_empty', ('has_language', 'fr')],
}
tests['lang_en'] = {
'matrix': {'query': 'paris', 'lang': 'en'},
- 'result_container': ['not_empty', ('has_lang', 'en')],
+ 'result_container': ['not_empty', ('has_language', 'en')],
}
if getattr(self.engine, 'safesearch', False):
diff --git a/searx/settings.yml b/searx/settings.yml
index 767bf6d82..d7149ad7c 100644
--- a/searx/settings.yml
+++ b/searx/settings.yml
@@ -105,11 +105,17 @@ outgoing: # communication with search engines
checker:
# disable checker when in debug mode
off_when_debug: True
+
# scheduling: interval or int
# use "scheduling: False" to disable scheduling
- scheduling:
- start_after: [300, 1800] # delay to start the first run of the checker
- every: [86400, 90000] # how often the checker runs
+ # to activate the scheduler:
+ # * uncomment "scheduling" section
+ # * add "cache2 = name=searxcache,items=2000,blocks=2000,blocksize=4096,bitmap=1" to your uwsgi.ini
+
+ # scheduling:
+ # start_after: [300, 1800] # delay to start the first run of the checker
+ # every: [86400, 90000] # how often the checker runs
+
# additional tests: only for the YAML anchors (see the engines section)
additional_tests:
rosebud: &test_rosebud
diff --git a/utils/lxc.sh b/utils/lxc.sh
index 39701ef04..bad5c10c8 100755
--- a/utils/lxc.sh
+++ b/utils/lxc.sh
@@ -39,6 +39,7 @@ $ubu1904_boilerplate
echo 'Set disable_coredump false' >> /etc/sudo.conf
"
+# shellcheck disable=SC2034
ubu2010_boilerplate="$ubu1904_boilerplate"
# shellcheck disable=SC2034
diff --git a/utils/searx.sh b/utils/searx.sh
index f85935fa2..a33642ee4 100755
--- a/utils/searx.sh
+++ b/utils/searx.sh
@@ -46,7 +46,6 @@ SEARX_PACKAGES_debian="\
python3-dev python3-babel python3-venv
uwsgi uwsgi-plugin-python3
git build-essential libxslt-dev zlib1g-dev libffi-dev libssl-dev
-libprotobuf-dev protobuf-compiler
shellcheck"
BUILD_PACKAGES_debian="\
@@ -59,7 +58,6 @@ SEARX_PACKAGES_arch="\
python python-pip python-lxml python-babel
uwsgi uwsgi-plugin-python
git base-devel libxml2
-protobuf
shellcheck"
BUILD_PACKAGES_arch="\
@@ -71,7 +69,7 @@ SEARX_PACKAGES_fedora="\
python python-pip python-lxml python-babel
uwsgi uwsgi-plugin-python3
git @development-tools libxml2
-ShellCheck protobuf-compiler protobuf-devel"
+ShellCheck"
BUILD_PACKAGES_fedora="\
firefox graphviz graphviz-gd ImageMagick librsvg2-tools
@@ -84,7 +82,7 @@ SEARX_PACKAGES_centos="\
python36 python36-pip python36-lxml python-babel
uwsgi uwsgi-plugin-python3
git @development-tools libxml2
-ShellCheck protobuf-compiler protobuf-devel"
+ShellCheck"
BUILD_PACKAGES_centos="\
firefox graphviz graphviz-gd ImageMagick librsvg2-tools