2021-01-13 11:31:25 +01:00
|
|
|
# SPDX-License-Identifier: AGPL-3.0-or-later
|
2024-03-11 07:45:08 +01:00
|
|
|
"""The JSON engine is a *generic* engine with which it is possible to configure
|
|
|
|
engines in the settings.
|
|
|
|
|
2025-01-11 01:51:33 +01:00
|
|
|
Configuration
|
|
|
|
=============
|
2024-03-11 07:45:08 +01:00
|
|
|
|
2025-01-11 01:51:33 +01:00
|
|
|
Request:
|
2024-03-11 07:45:08 +01:00
|
|
|
|
2025-01-11 01:51:33 +01:00
|
|
|
- :py:obj:`search_url`
|
2025-01-11 02:25:40 +01:00
|
|
|
- :py:obj:`lang_all`
|
|
|
|
- :py:obj:`soft_max_redirects`
|
2025-01-11 01:51:33 +01:00
|
|
|
- :py:obj:`method`
|
|
|
|
- :py:obj:`request_body`
|
|
|
|
- :py:obj:`cookies`
|
|
|
|
- :py:obj:`headers`
|
|
|
|
|
|
|
|
Paging:
|
|
|
|
|
|
|
|
- :py:obj:`paging`
|
|
|
|
- :py:obj:`page_size`
|
|
|
|
- :py:obj:`first_page_num`
|
|
|
|
|
2025-01-11 02:25:40 +01:00
|
|
|
Time Range:
|
|
|
|
|
|
|
|
- :py:obj:`time_range_support`
|
|
|
|
- :py:obj:`time_range_url`
|
|
|
|
- :py:obj:`time_range_map`
|
|
|
|
|
|
|
|
Safe-Search:
|
|
|
|
|
|
|
|
- :py:obj:`safe_search_support`
|
|
|
|
- :py:obj:`safe_search_map`
|
|
|
|
|
2025-01-11 01:51:33 +01:00
|
|
|
Response:
|
|
|
|
|
|
|
|
- :py:obj:`title_html_to_text`
|
|
|
|
- :py:obj:`content_html_to_text`
|
2025-01-11 02:25:40 +01:00
|
|
|
- :py:obj:`no_result_for_http_status`
|
2025-01-11 01:51:33 +01:00
|
|
|
|
|
|
|
JSON query:
|
|
|
|
|
|
|
|
- :py:obj:`results_query`
|
|
|
|
- :py:obj:`url_query`
|
|
|
|
- :py:obj:`url_prefix`
|
|
|
|
- :py:obj:`title_query`
|
|
|
|
- :py:obj:`content_query`
|
2025-01-11 02:25:40 +01:00
|
|
|
- :py:obj:`thumbnail_query`
|
|
|
|
- :py:obj:`thumbnail_prefix`
|
2025-01-11 01:51:33 +01:00
|
|
|
- :py:obj:`suggestion_query`
|
|
|
|
|
|
|
|
|
|
|
|
Example
|
|
|
|
=======
|
|
|
|
|
|
|
|
Here is a simple example of a JSON engine configure in the :ref:`settings
|
|
|
|
engine` section, further read :ref:`engines-dev`.
|
|
|
|
|
|
|
|
.. code:: yaml
|
|
|
|
|
|
|
|
- name : mdn
|
|
|
|
engine : json_engine
|
|
|
|
paging : True
|
|
|
|
search_url : https://developer.mozilla.org/api/v1/search?q={query}&page={pageno}
|
|
|
|
results_query : documents
|
|
|
|
url_query : mdn_url
|
|
|
|
url_prefix : https://developer.mozilla.org
|
|
|
|
title_query : title
|
|
|
|
content_query : summary
|
|
|
|
|
|
|
|
Implementations
|
|
|
|
===============
|
2024-03-11 07:45:08 +01:00
|
|
|
|
|
|
|
"""
|
2021-01-13 11:31:25 +01:00
|
|
|
|
2020-10-05 12:52:08 +02:00
|
|
|
from collections.abc import Iterable
|
2016-11-30 18:43:03 +01:00
|
|
|
from json import loads
|
2020-08-06 17:42:46 +02:00
|
|
|
from urllib.parse import urlencode
|
2021-02-10 16:40:03 +01:00
|
|
|
from searx.utils import to_string, html_to_text
|
2025-01-11 02:25:40 +01:00
|
|
|
from searx.network import raise_for_httperror
|
2016-11-30 18:43:03 +01:00
|
|
|
|
2014-01-20 02:31:20 +01:00
|
|
|
search_url = None
|
2025-01-11 01:51:33 +01:00
|
|
|
"""
|
|
|
|
Search URL of the engine. Example::
|
|
|
|
|
2025-01-11 02:25:40 +01:00
|
|
|
https://example.org/?search={query}&page={pageno}{time_range}{safe_search}
|
2025-01-11 01:51:33 +01:00
|
|
|
|
|
|
|
Replacements are:
|
|
|
|
|
|
|
|
``{query}``:
|
|
|
|
Search terms from user.
|
|
|
|
|
|
|
|
``{pageno}``:
|
|
|
|
Page number if engine supports paging :py:obj:`paging`
|
|
|
|
|
2025-01-11 02:25:40 +01:00
|
|
|
``{lang}``:
|
|
|
|
ISO 639-1 language code (en, de, fr ..)
|
|
|
|
|
|
|
|
``{time_range}``:
|
|
|
|
:py:obj:`URL parameter <time_range_url>` if engine :py:obj:`supports time
|
|
|
|
range <time_range_support>`. The value for the parameter is taken from
|
|
|
|
:py:obj:`time_range_map`.
|
|
|
|
|
|
|
|
``{safe_search}``:
|
|
|
|
Safe-search :py:obj:`URL parameter <safe_search_map>` if engine
|
|
|
|
:py:obj:`supports safe-search <safe_search_support>`. The ``{safe_search}``
|
|
|
|
replacement is taken from the :py:obj:`safes_search_map`. Filter results::
|
|
|
|
|
|
|
|
0: none, 1: moderate, 2:strict
|
|
|
|
|
|
|
|
If not supported, the URL parameter is an empty string.
|
|
|
|
|
2025-01-11 01:51:33 +01:00
|
|
|
"""
|
|
|
|
|
2025-01-11 02:25:40 +01:00
|
|
|
lang_all = 'en'
|
|
|
|
'''Replacement ``{lang}`` in :py:obj:`search_url` if language ``all`` is
|
|
|
|
selected.
|
|
|
|
'''
|
|
|
|
|
|
|
|
no_result_for_http_status = []
|
|
|
|
'''Return empty result for these HTTP status codes instead of throwing an error.
|
|
|
|
|
|
|
|
.. code:: yaml
|
|
|
|
|
|
|
|
no_result_for_http_status: []
|
|
|
|
'''
|
|
|
|
|
|
|
|
soft_max_redirects = 0
|
|
|
|
'''Maximum redirects, soft limit. Record an error but don't stop the engine'''
|
|
|
|
|
2024-11-27 14:13:23 +01:00
|
|
|
method = 'GET'
|
2025-01-11 01:51:33 +01:00
|
|
|
'''Some engines might require to do POST requests for search.'''
|
|
|
|
|
2024-11-27 14:13:23 +01:00
|
|
|
request_body = ''
|
2025-01-11 01:51:33 +01:00
|
|
|
'''The body of the request. This can only be used if different :py:obj:`method`
|
|
|
|
is set, e.g. ``POST``. For formatting see the documentation of :py:obj:`search_url`.
|
|
|
|
|
|
|
|
Note: Curly brackets which aren't encapsulating a replacement placeholder
|
|
|
|
must be escaped by doubling each ``{`` and ``}``.
|
|
|
|
|
|
|
|
.. code:: yaml
|
|
|
|
|
|
|
|
request_body: >-
|
|
|
|
{{
|
|
|
|
"search": "{query}",
|
|
|
|
"page": {pageno},
|
|
|
|
"extra": {{
|
|
|
|
"time_range": {time_range},
|
|
|
|
"rating": "{safe_search}"
|
|
|
|
}}
|
|
|
|
}}
|
|
|
|
'''
|
2013-11-19 15:49:52 +01:00
|
|
|
|
2022-06-10 23:26:55 +02:00
|
|
|
cookies = {}
|
2025-01-11 01:51:33 +01:00
|
|
|
'''Some engines might offer different result based on cookies.
|
|
|
|
Possible use-case: To set safesearch cookie.'''
|
|
|
|
|
2022-06-10 23:26:55 +02:00
|
|
|
headers = {}
|
|
|
|
'''Some engines might offer different result based on cookies or headers.
|
|
|
|
Possible use-case: To set safesearch cookie or header to moderate.'''
|
|
|
|
|
2024-11-27 14:13:23 +01:00
|
|
|
paging = False
|
2025-01-11 01:51:33 +01:00
|
|
|
'''Engine supports paging [True or False].'''
|
|
|
|
|
2016-07-16 19:26:29 +10:00
|
|
|
page_size = 1
|
2025-01-11 01:51:33 +01:00
|
|
|
'''Number of results on each page. Only needed if the site requires not a page
|
|
|
|
number, but an offset.'''
|
|
|
|
|
2016-07-16 19:26:29 +10:00
|
|
|
first_page_num = 1
|
2025-01-11 01:51:33 +01:00
|
|
|
'''Number of the first page (usually 0 or 1).'''
|
2016-07-16 19:26:29 +10:00
|
|
|
|
2024-11-27 14:13:23 +01:00
|
|
|
results_query = ''
|
2025-01-11 01:51:33 +01:00
|
|
|
'''JSON query for the list of result items.
|
|
|
|
|
|
|
|
The query string is a slash `/` separated path of JSON key names.
|
|
|
|
Array entries can be specified using the index or can be omitted entirely,
|
|
|
|
in which case each entry is considered -
|
|
|
|
most implementations will default to the first entry in this case.
|
|
|
|
'''
|
|
|
|
|
2024-11-27 14:13:23 +01:00
|
|
|
url_query = None
|
2025-01-11 01:51:33 +01:00
|
|
|
'''JSON query of result's ``url``. For the query string documentation see :py:obj:`results_query`'''
|
|
|
|
|
2024-11-27 14:13:23 +01:00
|
|
|
url_prefix = ""
|
2025-01-11 01:51:33 +01:00
|
|
|
'''String to prepend to the result's ``url``.'''
|
|
|
|
|
2024-11-27 14:13:23 +01:00
|
|
|
title_query = None
|
2025-01-11 01:51:33 +01:00
|
|
|
'''JSON query of result's ``title``. For the query string documentation see :py:obj:`results_query`'''
|
|
|
|
|
2024-11-27 14:13:23 +01:00
|
|
|
content_query = None
|
2025-01-11 01:51:33 +01:00
|
|
|
'''JSON query of result's ``content``. For the query string documentation see :py:obj:`results_query`'''
|
|
|
|
|
2025-01-11 02:25:40 +01:00
|
|
|
thumbnail_query = False
|
|
|
|
'''JSON query of result's ``thumbnail``. For the query string documentation see :py:obj:`results_query`'''
|
|
|
|
|
|
|
|
thumbnail_prefix = ''
|
|
|
|
'''String to prepend to the result's ``thumbnail``.'''
|
|
|
|
|
2024-11-27 14:13:23 +01:00
|
|
|
suggestion_query = ''
|
2025-01-11 01:51:33 +01:00
|
|
|
'''JSON query of result's ``suggestion``. For the query string documentation see :py:obj:`results_query`'''
|
|
|
|
|
2024-11-27 14:13:23 +01:00
|
|
|
title_html_to_text = False
|
2025-01-11 01:51:33 +01:00
|
|
|
'''Extract text from a HTML title string'''
|
|
|
|
|
2024-11-27 14:13:23 +01:00
|
|
|
content_html_to_text = False
|
2025-01-11 01:51:33 +01:00
|
|
|
'''Extract text from a HTML content string'''
|
2024-11-27 14:13:23 +01:00
|
|
|
|
2025-01-11 02:25:40 +01:00
|
|
|
time_range_support = False
|
|
|
|
'''Engine supports search time range.'''
|
|
|
|
|
|
|
|
time_range_url = '&hours={time_range_val}'
|
|
|
|
'''Time range URL parameter in the in :py:obj:`search_url`. If no time range is
|
|
|
|
requested by the user, the URL parameter is an empty string. The
|
|
|
|
``{time_range_val}`` replacement is taken from the :py:obj:`time_range_map`.
|
|
|
|
|
|
|
|
.. code:: yaml
|
|
|
|
|
|
|
|
time_range_url : '&days={time_range_val}'
|
|
|
|
'''
|
|
|
|
|
|
|
|
time_range_map = {
|
|
|
|
'day': 24,
|
|
|
|
'week': 24 * 7,
|
|
|
|
'month': 24 * 30,
|
|
|
|
'year': 24 * 365,
|
|
|
|
}
|
|
|
|
'''Maps time range value from user to ``{time_range_val}`` in
|
|
|
|
:py:obj:`time_range_url`.
|
|
|
|
|
|
|
|
.. code:: yaml
|
|
|
|
|
|
|
|
time_range_map:
|
|
|
|
day: 1
|
|
|
|
week: 7
|
|
|
|
month: 30
|
|
|
|
year: 365
|
|
|
|
'''
|
|
|
|
|
|
|
|
safe_search_support = False
|
|
|
|
'''Engine supports safe-search.'''
|
|
|
|
|
|
|
|
safe_search_map = {0: '&filter=none', 1: '&filter=moderate', 2: '&filter=strict'}
|
|
|
|
'''Maps safe-search value to ``{safe_search}`` in :py:obj:`search_url`.
|
|
|
|
|
|
|
|
.. code:: yaml
|
|
|
|
|
|
|
|
safesearch: true
|
|
|
|
safes_search_map:
|
|
|
|
0: '&filter=none'
|
|
|
|
1: '&filter=moderate'
|
|
|
|
2: '&filter=strict'
|
|
|
|
|
|
|
|
'''
|
|
|
|
|
2014-01-20 02:31:20 +01:00
|
|
|
|
2013-11-19 15:49:52 +01:00
|
|
|
def iterate(iterable):
|
2024-03-11 07:45:08 +01:00
|
|
|
if isinstance(iterable, dict):
|
|
|
|
items = iterable.items()
|
2013-11-19 15:49:52 +01:00
|
|
|
|
|
|
|
else:
|
2024-03-11 07:45:08 +01:00
|
|
|
items = enumerate(iterable)
|
|
|
|
for index, value in items:
|
2013-11-19 15:49:52 +01:00
|
|
|
yield str(index), value
|
|
|
|
|
2014-01-20 02:31:20 +01:00
|
|
|
|
2013-11-19 15:49:52 +01:00
|
|
|
def is_iterable(obj):
|
2024-03-11 07:45:08 +01:00
|
|
|
if isinstance(obj, str):
|
2014-01-20 02:31:20 +01:00
|
|
|
return False
|
2013-11-19 15:49:52 +01:00
|
|
|
return isinstance(obj, Iterable)
|
|
|
|
|
2014-01-20 02:31:20 +01:00
|
|
|
|
2024-03-11 07:45:08 +01:00
|
|
|
def parse(query): # pylint: disable=redefined-outer-name
|
|
|
|
q = [] # pylint: disable=invalid-name
|
2013-11-19 15:49:52 +01:00
|
|
|
for part in query.split('/'):
|
|
|
|
if part == '':
|
|
|
|
continue
|
2024-03-11 07:45:08 +01:00
|
|
|
q.append(part)
|
2013-11-19 15:49:52 +01:00
|
|
|
return q
|
|
|
|
|
2014-01-20 02:31:20 +01:00
|
|
|
|
2024-03-11 07:45:08 +01:00
|
|
|
def do_query(data, q): # pylint: disable=invalid-name
|
2013-11-19 15:49:52 +01:00
|
|
|
ret = []
|
2014-02-11 13:13:51 +01:00
|
|
|
if not q:
|
2013-11-19 15:49:52 +01:00
|
|
|
return ret
|
|
|
|
|
|
|
|
qkey = q[0]
|
|
|
|
|
2014-01-20 02:31:20 +01:00
|
|
|
for key, value in iterate(data):
|
2013-11-19 15:49:52 +01:00
|
|
|
|
|
|
|
if len(q) == 1:
|
|
|
|
if key == qkey:
|
|
|
|
ret.append(value)
|
|
|
|
elif is_iterable(value):
|
|
|
|
ret.extend(do_query(value, q))
|
|
|
|
else:
|
|
|
|
if not is_iterable(value):
|
|
|
|
continue
|
|
|
|
if key == qkey:
|
|
|
|
ret.extend(do_query(value, q[1:]))
|
|
|
|
else:
|
|
|
|
ret.extend(do_query(value, q))
|
|
|
|
return ret
|
|
|
|
|
2014-01-20 02:31:20 +01:00
|
|
|
|
2013-11-19 15:49:52 +01:00
|
|
|
def query(data, query_string):
|
|
|
|
q = parse(query_string)
|
|
|
|
|
|
|
|
return do_query(data, q)
|
|
|
|
|
2014-01-20 02:31:20 +01:00
|
|
|
|
2024-03-11 07:45:08 +01:00
|
|
|
def request(query, params): # pylint: disable=redefined-outer-name
|
2025-01-11 01:51:33 +01:00
|
|
|
'''Build request parameters (see :ref:`engine request`).'''
|
2025-01-11 02:25:40 +01:00
|
|
|
lang = lang_all
|
|
|
|
if params['language'] != 'all':
|
|
|
|
lang = params['language'][:2]
|
|
|
|
|
|
|
|
time_range = ''
|
|
|
|
if params.get('time_range'):
|
|
|
|
time_range_val = time_range_map.get(params.get('time_range'))
|
|
|
|
time_range = time_range_url.format(time_range_val=time_range_val)
|
|
|
|
|
|
|
|
safe_search = ''
|
|
|
|
if params['safesearch']:
|
|
|
|
safe_search = safe_search_map[params['safesearch']]
|
|
|
|
|
|
|
|
fp = { # pylint: disable=invalid-name
|
|
|
|
'query': urlencode({'q': query})[2:],
|
|
|
|
'lang': lang,
|
|
|
|
'pageno': (params['pageno'] - 1) * page_size + first_page_num,
|
|
|
|
'time_range': time_range,
|
|
|
|
'safe_search': safe_search,
|
|
|
|
}
|
2016-07-16 19:26:29 +10:00
|
|
|
|
2022-06-10 23:26:55 +02:00
|
|
|
params['cookies'].update(cookies)
|
|
|
|
params['headers'].update(headers)
|
|
|
|
|
2016-07-16 19:26:29 +10:00
|
|
|
params['url'] = search_url.format(**fp)
|
2024-11-27 14:13:23 +01:00
|
|
|
params['method'] = method
|
|
|
|
|
|
|
|
if request_body:
|
|
|
|
# don't url-encode the query if it's in the request body
|
|
|
|
fp['query'] = query
|
|
|
|
params['data'] = request_body.format(**fp)
|
2016-07-16 19:26:29 +10:00
|
|
|
|
2025-01-11 02:25:40 +01:00
|
|
|
params['soft_max_redirects'] = soft_max_redirects
|
|
|
|
params['raise_for_httperror'] = False
|
|
|
|
|
2013-11-19 15:49:52 +01:00
|
|
|
return params
|
|
|
|
|
|
|
|
|
2021-02-10 16:40:03 +01:00
|
|
|
def identity(arg):
|
|
|
|
return arg
|
|
|
|
|
|
|
|
|
2025-01-11 16:47:50 +01:00
|
|
|
def extract_response_info(result):
|
|
|
|
title_filter = html_to_text if title_html_to_text else identity
|
|
|
|
content_filter = html_to_text if content_html_to_text else identity
|
|
|
|
|
|
|
|
tmp_result = {}
|
|
|
|
|
|
|
|
try:
|
|
|
|
url = query(result, url_query)[0]
|
|
|
|
tmp_result['url'] = url_prefix + to_string(url)
|
|
|
|
|
|
|
|
title = query(result, title_query)[0]
|
|
|
|
tmp_result['title'] = title_filter(to_string(title))
|
|
|
|
except: # pylint: disable=bare-except
|
|
|
|
return None
|
|
|
|
|
|
|
|
try:
|
|
|
|
content = query(result, content_query)[0]
|
|
|
|
tmp_result['content'] = content_filter(to_string(content))
|
|
|
|
except: # pylint: disable=bare-except
|
|
|
|
tmp_result['content'] = ""
|
|
|
|
|
|
|
|
try:
|
|
|
|
if thumbnail_query:
|
|
|
|
thumbnail_query_result = query(result, thumbnail_query)[0]
|
|
|
|
tmp_result['thumbnail'] = thumbnail_prefix + to_string(thumbnail_query_result)
|
|
|
|
except: # pylint: disable=bare-except
|
|
|
|
pass
|
|
|
|
|
|
|
|
return tmp_result
|
|
|
|
|
|
|
|
|
2013-11-19 15:49:52 +01:00
|
|
|
def response(resp):
|
2025-01-11 01:51:33 +01:00
|
|
|
'''Scrap *results* from the response (see :ref:`engine results`).'''
|
2013-11-19 15:49:52 +01:00
|
|
|
results = []
|
2025-01-11 01:51:33 +01:00
|
|
|
|
2025-01-11 02:25:40 +01:00
|
|
|
if no_result_for_http_status and resp.status_code in no_result_for_http_status:
|
|
|
|
return results
|
|
|
|
|
|
|
|
raise_for_httperror(resp)
|
|
|
|
|
2025-01-11 01:51:33 +01:00
|
|
|
if not resp.text:
|
|
|
|
return results
|
|
|
|
|
2013-11-19 15:49:52 +01:00
|
|
|
json = loads(resp.text)
|
2025-01-11 02:25:40 +01:00
|
|
|
is_onion = 'onions' in categories
|
2021-02-10 16:40:03 +01:00
|
|
|
|
2016-08-12 20:15:27 +10:00
|
|
|
if results_query:
|
2024-03-11 07:45:08 +01:00
|
|
|
rs = query(json, results_query) # pylint: disable=invalid-name
|
|
|
|
if not rs:
|
2017-11-02 00:43:29 +01:00
|
|
|
return results
|
2025-01-11 16:47:50 +01:00
|
|
|
rs = rs[0] # pylint: disable=invalid-name
|
2016-08-12 20:15:27 +10:00
|
|
|
else:
|
2025-01-11 16:47:50 +01:00
|
|
|
rs = json # pylint: disable=invalid-name
|
|
|
|
|
|
|
|
for result in rs:
|
|
|
|
tmp_result = extract_response_info(result)
|
|
|
|
if not tmp_result:
|
|
|
|
continue
|
|
|
|
|
|
|
|
if is_onion:
|
|
|
|
tmp_result['is_onion'] = True
|
|
|
|
|
|
|
|
results.append(tmp_result)
|
2016-08-12 20:15:27 +10:00
|
|
|
|
|
|
|
if not suggestion_query:
|
|
|
|
return results
|
|
|
|
for suggestion in query(json, suggestion_query):
|
|
|
|
results.append({'suggestion': suggestion})
|
2013-11-19 15:49:52 +01:00
|
|
|
return results
|