From f596f5767bed915a5c3bed59ae26283e53f975ca Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?No=C3=A9mi=20V=C3=A1nyi?= Date: Sat, 27 Mar 2021 16:50:03 +0100 Subject: [PATCH] fix Semantic Scholar engine --- searx/engines/semantic_scholar.py | 43 +++++++++++++++++++++++++++++++ searx/settings.yml | 11 +++----- 2 files changed, 46 insertions(+), 8 deletions(-) create mode 100644 searx/engines/semantic_scholar.py diff --git a/searx/engines/semantic_scholar.py b/searx/engines/semantic_scholar.py new file mode 100644 index 00000000..a49f84b6 --- /dev/null +++ b/searx/engines/semantic_scholar.py @@ -0,0 +1,43 @@ +# SPDX-License-Identifier: AGPL-3.0-or-later +""" + Semantic Scholar (Science) +""" + +from json import dumps, loads + + +search_url = 'https://www.semanticscholar.org/api/1/search' + + +def request(query, params): + params['url'] = search_url + params['method'] = 'POST' + params['headers']['content-type'] = 'application/json' + params['data'] = dumps({ + "queryString": query, + "page": params['pageno'], + "pageSize": 10, + "sort": "relevance", + "useFallbackRankerService": False, + "useFallbackSearchCluster": False, + "performTitleMatch": True, + "getQuerySuggestions": False, + "authors": [], + "coAuthors": [], + "venues": [], + "performTitleMatch": True, + }) + return params + + +def response(resp): + res = loads(resp.text) + results = [] + for result in res['results']: + results.append({ + 'url': result['primaryPaperLink']['url'], + 'title': result['title']['text'], + 'content': result['paperAbstractTruncated'] + }) + + return results diff --git a/searx/settings.yml b/searx/settings.yml index 1958210d..84aca86f 100644 --- a/searx/settings.yml +++ b/searx/settings.yml @@ -918,13 +918,8 @@ engines: # disabled : True - name : semantic scholar - engine : xpath - paging : True - search_url : https://www.semanticscholar.org/search?q={query}&sort=relevance&page={pageno}&ae=false - results_xpath : //article - url_xpath : .//div[@class="search-result-title"]/a/@href - title_xpath : .//div[@class="search-result-title"]/a - content_xpath : .//div[@class="search-result-abstract"] + engine : semantic_scholar + disabled : True shortcut : se categories : science about: @@ -933,7 +928,7 @@ engines: official_api_documentation: https://api.semanticscholar.org/ use_official_api: false require_api_key: false - results: HTML + results: JSON # Spotify needs API credentials # - name : spotify