search

ryanmuller / Rsrchr

Platform for collaborative research
2 stars 0 forks source link

RefLinks export API #25

Open ryanmuller opened 12 years ago

ryanmuller commented 12 years ago

Get a list of other wiki links about a paper.

houshuang commented 12 years ago

In addition to a JSON api, make this a Javascript widget that can be embedded in blogs (other people who wrote about this paper), on researchr ref pages etc.

(Fancier would be a bookmarklet that used dotjs/greasemonkey to detect papers listed on Mendeley, GScholar etc and added links if articles are already on Scrobblr, how many external links etc)

ryanmuller commented 12 years ago

The new bookmarklet (pushing now) will do either "Import" or "View" depending on whether the citation page exists. Still only Google Scholar.

houshuang commented 12 years ago

First was confused because I thought it worked as before (select text, and press bookmarklet), clicked it a bunch of times and got a ton of "Add to Scrobblr" links :) Then I realized what it was doing. It's neat that it adds a link to all publications (very nice with dotjs I assume), however the links don't do anything on my computer (Chrome Dev).

houshuang commented 12 years ago

Also according to the code, it checks if a citation already exists, and shows "View" if it exists. However, when I run it on "Induction and Deduction in Bayesian Data Analysis", which you just scrobbled (and which has the same citekey in the system as in the Google BibTeX), it says "import", not "view".

ryanmuller commented 12 years ago

Are you sure the import links don't do anything? They don't change after you click on them, but you can check whether they've imported something e.g. by going to your user page: http://stormy-leaf-9036.herokuapp.com/users/3

I did try to get the links to change on success but that wasn't working for some reason. But I'll make an issue for that.

The view thing is working fine for me. I wonder if it's an error with the regex for that article? Try it searching for "koedinger"

2012/3/11 Stian Hklev < reply@reply.github.com

First was confused because I thought it worked as before (select text, and press bookmarklet), clicked it a bunch of times and got a ton of "Add to Scrobblr" links :) Then I realized what it was doing. It's neat that it adds a link to all publications (very nice with dotjs I assume), however the links don't do anything on my computer (Chrome Dev).

Reply to this email directly or view it on GitHub: https://github.com/cicatriz/Rsrchr/issues/25#issuecomment-4444820

houshuang commented 12 years ago

I wonder if it is because Google Scholar gets suspicious very quickly. When I try to click on a link to a bibtex page manually to check the citekey, I get a captcha every time... This is what I had problems with when I tried to automatize access to GScholar from BibDesk... We're not even talking scraping thousands of entries - even a few clicks rapidly, and GScholar shuts down. So it's possible that it quickly shuts out the javascript after just fetching a few BibTeX entries, and thus the rest of the functionality does not work.

In this case, the first version of your script might work much better - or one that doesn't prefetch all the BibTeX to check whether it exists or not.

I'll say it again - we desperately need an open version of GScholar... But no idea who would start working on that...

ryanmuller commented 12 years ago

Bizarre. Is that an outside US thing? You can see what response codes the calls are returning in the Network tab of JS Console (opt-cmd-J) -- they're the scholar.bib ones.

But yeah, I could just have it redirect to the existing one on the server if you try to submit a duplicate citekey. That means you'd have to cmd-click to import multiple, but at least you get a response either way.

2012/3/12 Stian Hklev < reply@reply.github.com

I wonder if it is because Google Scholar gets suspicious very quickly. When I try to click on a link to a bibtex page manually to check the citekey, I get a captcha every time... This is what I had problems with when I tried to automatize access to GScholar from BibDesk... We're not even talking scraping thousands of entries - even a few clicks rapidly, and GScholar shuts down. So it's possible that it quickly shuts out the javascript after just fetching a few BibTeX entries, and thus the rest of the functionality does not work.

In this case, the first version of your script might work much better - or one that doesn't prefetch all the BibTeX to check whether it exists or not.

I'll say it again - we desperately need an open version of GScholar... But no idea who would start working on that...

Reply to this email directly or view it on GitHub: https://github.com/cicatriz/Rsrchr/issues/25#issuecomment-4450509