summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
-rw-r--r--Makefile15
-rw-r--r--docs/dev/engine_overview.rst2
-rw-r--r--requirements-dev.txt4
-rw-r--r--searx/engines/__init__.py33
-rw-r--r--searx/engines/apkmirror.py42
-rw-r--r--searx/engines/solidtorrents.py53
-rw-r--r--searx/poolrequests.py16
-rw-r--r--searx/search/processors/online.py2
-rw-r--r--searx/settings.yml7
-rw-r--r--searx/templates/oscar/preferences.html2
-rw-r--r--searx/templates/simple/preferences.html2
11 files changed, 113 insertions, 65 deletions
diff --git a/Makefile b/Makefile
index 41d7c7f06..a4e9110ed 100644
--- a/Makefile
+++ b/Makefile
@@ -94,15 +94,14 @@ project: buildenv useragents.update engines.languages
engines.languages: pyenvinstall
$(Q)echo "fetch languages .."
- $(Q)$(PY_ENV_ACT); python utils/fetch_languages.py
- $(Q)echo "update searx/data/engines_languages.json"
- $(Q)mv engines_languages.json searx/data/engines_languages.json
- $(Q)echo "update searx/languages.py"
- $(Q)mv languages.py searx/languages.py
+ $(Q)$(PY_ENV_ACT); python ./searx_extra/update/update_languages.py
+ $(Q)echo "updated searx/data/engines_languages.json"
+ $(Q)echo "updated searx/languages.py"
useragents.update: pyenvinstall
- $(Q)echo "Update searx/data/useragents.json with the most recent versions of Firefox."
- $(Q)$(PY_ENV_ACT); python utils/fetch_firefox_version.py
+ $(Q)echo "fetch useragents .."
+ $(Q)$(PY_ENV_ACT); python ./searx_extra/update/update_firefox_version.py
+ $(Q)echo "updated searx/data/useragents.json with the most recent versions of Firefox."
buildenv: pyenv
$(Q)$(PY_ENV_ACT); SEARX_DEBUG=1 python utils/build_env.py
@@ -192,8 +191,10 @@ PYLINT_FILES=\
searx/engines/google_videos.py \
searx/engines/google_images.py \
searx/engines/mediathekviewweb.py \
+ searx/engines/solidtorrents.py \
searx/engines/google_scholar.py \
searx/engines/yahoo_news.py \
+ searx/engines/apkmirror.py \
searx_extra/update/update_external_bangs.py
test.pylint: pyenvinstall
diff --git a/docs/dev/engine_overview.rst b/docs/dev/engine_overview.rst
index b51181974..42c205d9d 100644
--- a/docs/dev/engine_overview.rst
+++ b/docs/dev/engine_overview.rst
@@ -58,6 +58,8 @@ argument type information
name string name of search-engine
engine string name of searx-engine
(filename without ``.py``)
+enable_http bool enable HTTP
+ (by default only HTTPS is enabled).
shortcut string shortcut of search-engine
timeout string specific timeout for search-engine
display_error_messages boolean display error messages on the web UI
diff --git a/requirements-dev.txt b/requirements-dev.txt
index c3cd35432..44093f216 100644
--- a/requirements-dev.txt
+++ b/requirements-dev.txt
@@ -8,10 +8,10 @@ transifex-client==0.14.2
selenium==3.141.0
twine==3.3.0
Pallets-Sphinx-Themes==1.2.3
-Sphinx==3.5.1
+Sphinx==3.5.2
sphinx-issues==1.2.0
sphinx-jinja==1.1.1
-sphinx-tabs==2.0.1
+sphinx-tabs==2.1.0
sphinxcontrib-programoutput==0.16
sphinx-autobuild==2020.9.1
linuxdoc==20210110
diff --git a/searx/engines/__init__.py b/searx/engines/__init__.py
index 7270724b6..2238ea1b9 100644
--- a/searx/engines/__init__.py
+++ b/searx/engines/__init__.py
@@ -50,6 +50,7 @@ engine_default_args = {'paging': False,
'timeout': settings['outgoing']['request_timeout'],
'shortcut': '-',
'disabled': False,
+ 'enable_http': False,
'suspend_end_time': 0,
'continuous_errors': 0,
'time_range_support': False,
@@ -305,35 +306,3 @@ def initialize_engines(engine_list):
if init_fn:
logger.debug('%s engine: Starting background initialization', engine_name)
threading.Thread(target=engine_init, args=(engine_name, init_fn)).start()
-
- _set_https_support_for_engine(engine)
-
-
-def _set_https_support_for_engine(engine):
- # check HTTPS support if it is not disabled
- if engine.engine_type != 'offline' and not hasattr(engine, 'https_support'):
- params = engine.request('http_test', {
- 'method': 'GET',
- 'headers': {},
- 'data': {},
- 'url': '',
- 'cookies': {},
- 'verify': True,
- 'auth': None,
- 'pageno': 1,
- 'time_range': None,
- 'language': '',
- 'safesearch': False,
- 'is_test': True,
- 'category': 'files',
- 'raise_for_status': True,
- 'engine_data': {},
- })
-
- if 'url' not in params:
- return
-
- parsed_url = urlparse(params['url'])
- https_support = parsed_url.scheme == 'https'
-
- setattr(engine, 'https_support', https_support)
diff --git a/searx/engines/apkmirror.py b/searx/engines/apkmirror.py
index a9ddd711a..05a635883 100644
--- a/searx/engines/apkmirror.py
+++ b/searx/engines/apkmirror.py
@@ -1,13 +1,21 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
+"""APKMirror
"""
- APK Mirror
-"""
+
+# pylint: disable=invalid-name, missing-function-docstring
from urllib.parse import urlencode
from lxml import html
-from searx.utils import extract_text, eval_xpath_list, eval_xpath_getindex
-# about
+from searx import logger
+from searx.utils import (
+ eval_xpath_list,
+ eval_xpath_getindex,
+ extract_text,
+)
+
+logger = logger.getChild('APKMirror engine')
+
about = {
"website": 'https://www.apkmirror.com',
"wikidata_id": None,
@@ -18,11 +26,8 @@ about = {
}
# engine dependent config
-categories = ['it']
+categories = ['files']
paging = True
-
-# I am not 100% certain about this, as apkmirror appears to be a wordpress site,
-# which might support time_range searching. If you want to implement it, go ahead.
time_range_support = False
# search-url
@@ -30,37 +35,34 @@ base_url = 'https://www.apkmirror.com'
search_url = base_url + '/?post_type=app_release&searchtype=apk&page={pageno}&{query}'
-# do search-request
def request(query, params):
-
- params['url'] = search_url.format(pageno=params['pageno'],
- query=urlencode({'s': query}))
+ params['url'] = search_url.format(
+ pageno = params['pageno'],
+ query = urlencode({'s': query}),
+ )
+ logger.debug("query_url --> %s", params['url'])
return params
-# get response from search-request
def response(resp):
results = []
dom = html.fromstring(resp.text)
# parse results
- for result in eval_xpath_list(dom, './/div[@id="content"]/div[@class="listWidget"]//div[@class="appRow"]'):
+ for result in eval_xpath_list(dom, "//div[@id='content']//div[@class='listWidget']/div/div[@class='appRow']"):
link = eval_xpath_getindex(result, './/h5/a', 0)
+
url = base_url + link.attrib.get('href') + '#downloads'
title = extract_text(link)
- thumbnail_src = base_url\
- + eval_xpath_getindex(result, './/img', 0).attrib.get('src').replace('&w=32&h=32', '&w=64&h=64')
-
+ img_src = base_url + eval_xpath_getindex(result, './/img/@src', 0)
res = {
'url': url,
'title': title,
- 'thumbnail_src': thumbnail_src
+ 'img_src': img_src
}
- # append result
results.append(res)
- # return results
return results
diff --git a/searx/engines/solidtorrents.py b/searx/engines/solidtorrents.py
new file mode 100644
index 000000000..050149187
--- /dev/null
+++ b/searx/engines/solidtorrents.py
@@ -0,0 +1,53 @@
+# SPDX-License-Identifier: AGPL-3.0-or-later
+"""Solid Torrents
+
+"""
+
+# pylint: disable=missing-function-docstring
+
+from json import loads
+from urllib.parse import urlencode
+from searx import logger
+
+logger = logger.getChild('solidtor engine')
+
+about = {
+ "website": 'https://www.solidtorrents.net/',
+ "wikidata_id": None,
+ "official_api_documentation": None,
+ "use_official_api": True,
+ "require_api_key": False,
+ "results": 'JSON',
+}
+
+categories = ['files']
+paging = True
+
+base_url = 'https://www.solidtorrents.net/'
+search_url = base_url + 'api/v1/search?{query}'
+
+
+def request(query, params):
+ skip = (params['pageno'] - 1) * 20
+ query = urlencode({'q': query, 'skip': skip})
+ params['url'] = search_url.format(query=query)
+ logger.debug("query_url --> %s", params['url'])
+ return params
+
+
+def response(resp):
+ results = []
+ search_results = loads(resp.text)
+
+ for result in search_results["results"]:
+ results.append({
+ 'infohash': result["infohash"],
+ 'seed': result["swarm"]["seeders"],
+ 'leech': result["swarm"]["leechers"],
+ 'title': result["title"],
+ 'link': "https://solidtorrents.net/view/" + result["_id"],
+ 'filesize': result["size"],
+ 'magnetlink': result["magnet"],
+ 'template': "torrent.html",
+ })
+ return results
diff --git a/searx/poolrequests.py b/searx/poolrequests.py
index 8b8681437..ab327251b 100644
--- a/searx/poolrequests.py
+++ b/searx/poolrequests.py
@@ -91,9 +91,10 @@ class SessionSinglePool(requests.Session):
self.adapters.clear()
https_adapter = threadLocal.__dict__.setdefault('https_adapter', next(https_adapters))
- http_adapter = threadLocal.__dict__.setdefault('http_adapter', next(http_adapters))
self.mount('https://', https_adapter)
- self.mount('http://', http_adapter)
+ if get_enable_http_protocol():
+ http_adapter = threadLocal.__dict__.setdefault('http_adapter', next(http_adapters))
+ self.mount('http://', http_adapter)
def close(self):
"""Call super, but clear adapters since there are managed globaly"""
@@ -106,6 +107,17 @@ def set_timeout_for_thread(timeout, start_time=None):
threadLocal.start_time = start_time
+def set_enable_http_protocol(enable_http):
+ threadLocal.enable_http = enable_http
+
+
+def get_enable_http_protocol():
+ try:
+ return threadLocal.enable_http
+ except AttributeError:
+ return False
+
+
def reset_time_for_thread():
threadLocal.total_time = 0
diff --git a/searx/search/processors/online.py b/searx/search/processors/online.py
index df0ab8c21..1fc6444ad 100644
--- a/searx/search/processors/online.py
+++ b/searx/search/processors/online.py
@@ -131,6 +131,8 @@ class OnlineProcessor(EngineProcessor):
poolrequests.set_timeout_for_thread(timeout_limit, start_time=start_time)
# reset the HTTP total time
poolrequests.reset_time_for_thread()
+ # enable HTTP only if explicitly enabled
+ poolrequests.set_enable_http_protocol(self.engine.enable_http)
# suppose everything will be alright
requests_exception = False
diff --git a/searx/settings.yml b/searx/settings.yml
index e45afb59b..85ba4b2fe 100644
--- a/searx/settings.yml
+++ b/searx/settings.yml
@@ -656,6 +656,7 @@ engines:
- name : library genesis
engine : xpath
+ enable_http: True
search_url : http://libgen.rs/search.php?req={query}
url_xpath : //a[contains(@href,"bookfi.net/md5")]/@href
title_xpath : //a[contains(@href,"book/")]/text()[1]
@@ -956,6 +957,12 @@ engines:
timeout : 6.0
disabled : True
+ - name : solidtorrents
+ engine : solidtorrents
+ shortcut : solid
+ timeout : 3.0
+ disabled : True
+
- name : torrentz
engine : torrentz
shortcut : tor
diff --git a/searx/templates/oscar/preferences.html b/searx/templates/oscar/preferences.html
index fc20b8ca5..6253b9858 100644
--- a/searx/templates/oscar/preferences.html
+++ b/searx/templates/oscar/preferences.html
@@ -230,7 +230,7 @@
<td class="onoff-checkbox">
{{ checkbox_toggle('engine_' + search_engine.name|replace(' ', '_') + '__' + categ|replace(' ', '_'), (search_engine.name, categ) in disabled_engines) }}
</td>
- <th scope="row">{% if not search_engine.https_support %}{{ icon('exclamation-sign', 'No HTTPS') }}{% endif %} {{ search_engine.name }}</td></th>
+ <th scope="row">{% if search_engine.enable_http %}{{ icon('exclamation-sign', 'No HTTPS') }}{% endif %} {{ search_engine.name }}</td></th>
<td class="name">{{ shortcuts[search_engine.name] }}
<td>{{ support_toggle(stats[search_engine.name].supports_selected_language) }}</td>
<td>{{ support_toggle(search_engine.safesearch==True) }}</td>
diff --git a/searx/templates/simple/preferences.html b/searx/templates/simple/preferences.html
index f091a97cf..dff7ffba6 100644
--- a/searx/templates/simple/preferences.html
+++ b/searx/templates/simple/preferences.html
@@ -121,7 +121,7 @@
{% set engine_id = 'engine_' + search_engine.name|replace(' ', '_') + '__' + categ|replace(' ', '_') %}
<tr>
<td class="engine_checkbox">{{ checkbox_onoff(engine_id, (search_engine.name, categ) in disabled_engines) }}</td>
- <th class="name">{% if not search_engine.https_support %}{{ icon('warning', 'No HTTPS') }}{% endif %} {{ search_engine.name }}</th>
+ <th class="name">{% if search_engine.enable_http %}{{ icon('warning', 'No HTTPS') }}{% endif %} {{ search_engine.name }}</th>
<td class="shortcut">{{ shortcuts[search_engine.name] }}</td>
<td>{{ checkbox(engine_id + '_supported_languages', current_language == 'all' or current_language in search_engine.supported_languages or current_language.split('-')[0] in search_engine.supported_languages, true, true) }}</td>
<td>{{ checkbox(engine_id + '_safesearch', search_engine.safesearch==True, true, true) }}</td>