semanticscholar
semanticscholar copied to clipboard
Searching with keywords
Can we have a search mechanism for searching papers based on some keyword queries? There is already an API call for that: https://api.semanticscholar.org/graph/v1/paper/search?query=word+embedding+automatic+query+expansion&fields=title,abstract
I've been working to deliver a new release based on the recent Semantic Scholar API upgrade, which includes searching for papers by keyword.
Hi, I'm looking forward to this feature. Any chance you've gotten recently to wrap this feature up?
Hi, @wookayin.
Unfortunately, I've been overloaded the last few months with other projects and haven't found enough time to work on this one.
But I have some code already done and will give my best to release a new version that enables the search feature soon. Maybe until next week.
Most of the work on this is probably figuring out the best API within this package. In the meanwhile, I've been using this little function to get the paper/search functionality:
from urllib.parse import quote
import urllib.request
import json
def paper_search(query, limit=10):
query = quote(query)
url_base = f"https://api.semanticscholar.org/graph/v1/paper/search?limit={limit}&fields=title,authors&query="
url = url_base + query
data = urllib.request.urlopen(url)
data = data.read().decode("utf-8")
return json.loads(data)["data"]
result = paper_search("Backtranslation Feedback Improves User Confidence")
for line in result:
print(line)
Output:
{'paperId': 'f3c9084cd61dfcdc51c9d846a587863c17475540', 'title': 'Backtranslation Feedback Improves User Confidence in MT, Not Quality', 'authors': [{'authorId': '1429837660', 'name': 'Vilém Zouhar'}, {'authorId': '2072797005', 'name': "Michal Nov'ak"}, {'authorId': '1739254941', 'name': "Mat'uvs vZilinec"}, {'authorId': '151158933', 'name': 'Ondvrej Bojar'}, {'authorId': '2072784279', 'name': "Mateo Obreg'on"}, {'authorId': '3252072', 'name': 'Robin L. Hill'}, {'authorId': '1948646', 'name': 'F. Blain'}, {'authorId': '2006017', 'name': 'M. Fomicheva'}, {'authorId': '2065700826', 'name': 'Lucia Specia'}, {'authorId': '78447404', 'name': 'Lisa Yankovskaya'}]}
{'paperId': '4d9a4b4ac353a999c4fba3925739c9ea226b2af1', 'title': 'Optimally-calibrated non-invasive feedback improves amputees’ metabolic consumption, balance and walking confidence', 'authors': [{'authorId': '2146108553', 'name': 'Lauren Chee'}, {'authorId': '143696395', 'name': 'G. Valle'}, {'authorId': '2179503901', 'name': 'Michele Marazzi'}, {'authorId': '103705041', 'name': 'G. Preatoni'}, {'authorId': '151407570', 'name': 'F. Haufe'}, {'authorId': '144114747', 'name': 'M. Xiloyannis'}, {'authorId': '3134357', 'name': 'R. Riener'}, {'authorId': '2008652', 'name': 'S. Raspopovic'}]}
{'paperId': '4106abc70ced2ff666e94ea9b668128b0f982cce', 'title': 'High-Resolution Image Inpainting with Iterative Confidence Feedback and Guided Upsampling', 'authors': [{'authorId': '2111187011', 'name': 'Yu Zeng'}, {'authorId': '145527707', 'name': 'Zhe L. Lin'}, {'authorId': '1768964', 'name': 'Jimei Yang'}, {'authorId': '1519062024', 'name': 'Jianming Zhang'}, {'authorId': '2177801', 'name': 'E. Shechtman'}, {'authorId': '153176123', 'name': 'Huchuan Lu'}]}
...
I think your viewpoint is correct on that @zouharvi . I am currently working on how to deal with paginated results. I intend to allow the user to iterate over results regardless number of pages.
After a long delay, the new version was released and the search feature is finally here. I hope it could be helpful.