Merge pull request #944 from return42/fix-939

Browse Source 
[fix] engine: Semantic Scholar (Science) // rework & fix
This commit is contained in:
Markus Heiser 2022-03-05 11:58:14 +01:00 committed by GitHub
commit 0d86e7e4ea
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 44 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

53
searx/engines/semantic_scholar.py
View File

@ -1,12 +1,23 @@
# SPDX-License-Identifier: AGPL-3.0-or-later # SPDX-License-Identifier: AGPL-3.0-or-later
""" # lint: pylint
Semantic Scholar (Science) """Semantic Scholar (Science)
""" """
from json import dumps, loads from json import dumps, loads
from datetime import datetime
about = {
"website": 'https://www.semanticscholar.org/',
"wikidata_id": 'Q22908627',
"official_api_documentation": 'https://api.semanticscholar.org/',
"use_official_api": True,
"require_api_key": False,
"results": 'JSON',
}
paging = True
search_url = 'https://www.semanticscholar.org/api/1/search' search_url = 'https://www.semanticscholar.org/api/1/search'
paper_url = 'https://www.semanticscholar.org/paper'
def request(query, params): def request(query, params):
@ -34,13 +45,37 @@ def request(query, params):
def response(resp): def response(resp):
res = loads(resp.text) res = loads(resp.text)
results = [] results = []
for result in res['results']: for result in res['results']:
results.append( item = {}
{ metadata = []
'url': result['primaryPaperLink']['url'],
'title': result['title']['text'], url = result.get('primaryPaperLink', {}).get('url')
'content': result['paperAbstractTruncated'], if not url and result.get('links'):
} url = result.get('links')[0]
) if not url:
alternatePaperLinks = result.get('alternatePaperLinks')
if alternatePaperLinks:
url = alternatePaperLinks[0].get('url')
if not url:
url = paper_url + '/%s' % result['id']
item['url'] = url
item['title'] = result['title']['text']
item['content'] = result['paperAbstract']['text']
metadata = result.get('fieldsOfStudy') or []
venue = result.get('venue', {}).get('text')
if venue:
metadata.append(venue)
if metadata:
item['metadata'] = ', '.join(metadata)
pubDate = result.get('pubDate')
if pubDate:
item['publishedDate'] = datetime.strptime(pubDate, "%Y-%m-%d")
results.append(item)
return results return results

7
searx/settings.yml
View File

@ -1205,13 +1205,6 @@ engines:
disabled: true disabled: true
shortcut: se shortcut: se
categories: science categories: science
about:
website: https://www.semanticscholar.org/
wikidata_id: Q22908627
official_api_documentation: https://api.semanticscholar.org/
use_official_api: false
require_api_key: false
results: JSON
# Spotify needs API credentials # Spotify needs API credentials
# - name: spotify # - name: spotify