Re: [CODE4LIB] vivosearchlight
I haven't looked at my C4L messages in quite awhile so I missed this. I've asked him before about making the code available but at the time he didn't think it was ready to be shared. We've talked about some other applications for it and one of the things we discussed was making it work with Solr instead of elastic search. I'm going to be on a skype call with him next week for another project we're working on and can ask him more. I've also been looking at some text mining/entity mining apis (anyone else looking at AlchemyAPI?) that might improve the relevancy of the search results but the result are only going to be as good as what's in the search index and I haven't really looked at that piece. -Original Message- From: Code for Libraries [mailto:CODE4LIB@LISTSERV.ND.EDU] On Behalf Of Ed Summers Sent: Sunday, November 20, 2011 6:54 AM To: CODE4LIB@LISTSERV.ND.EDU Subject: [CODE4LIB] vivosearchlight On Tue, Nov 1, 2011 at 7:44 AM, John Fereira ja...@cornell.edu wrote: If you want to see what node.js can do to implement a search mechanism take a look something one of my colleagues developed. http://vivosearchlight.org It installs a bookmarklet in your browser (take about 5 seconds) that will initiate a search against a solr index that contains user profile information from several institutions using VIVO (a semantic web application). From any web page, clicking on the Vivo Searchlight button in your browser will initiate a search and find experts with expertise relevant to the content of the page. Highlight some text on the page and it will re-execute a search with just those words. Thanks for sharing John. That's a really a neat idea, even if the results don't seem particularly relevant for some tests I tried. I was curious how it does the matching of page text against the profiles. I see from the description at http://vivosearchlight.org that EleasticSearch is being used instead of Solr. Any chance Miles Worthington (ok I googled) would be willing to share the source code on his github account [1], or elsewhere? //Ed [1] https://github.com/milesworthington
Re: [CODE4LIB] vivosearchlight
On Sun, Nov 20, 2011 at 6:53 AM, Ed Summers e...@pobox.com wrote: On Tue, Nov 1, 2011 at 7:44 AM, John Fereira ja...@cornell.edu wrote: If you want to see what node.js can do to implement a search mechanism take a look something one of my colleagues developed. http://vivosearchlight.org It installs a bookmarklet in your browser (take about 5 seconds) that will initiate a search against a solr index that contains user profile information from several institutions using VIVO (a semantic web application). From any web page, clicking on the Vivo Searchlight button in your browser will initiate a search and find experts with expertise relevant to the content of the page. Highlight some text on the page and it will re-execute a search with just those words. Thanks for sharing John. That's a really a neat idea, even if the results don't seem particularly relevant for some tests I tried. I was curious how it does the matching of page text against the profiles. I see from the description at http://vivosearchlight.org that EleasticSearch is being used instead of Solr. Any chance Miles Worthington (ok I googled) would be willing to share the source code on his github account [1], or elsewhere? //Ed I second the request. I've been doing a fair amount of work with Node and ElasticSearch as well lately. [1] https://github.com/milesworthington
[CODE4LIB] vivosearchlight
On Tue, Nov 1, 2011 at 7:44 AM, John Fereira ja...@cornell.edu wrote: If you want to see what node.js can do to implement a search mechanism take a look something one of my colleagues developed. http://vivosearchlight.org It installs a bookmarklet in your browser (take about 5 seconds) that will initiate a search against a solr index that contains user profile information from several institutions using VIVO (a semantic web application). From any web page, clicking on the Vivo Searchlight button in your browser will initiate a search and find experts with expertise relevant to the content of the page. Highlight some text on the page and it will re-execute a search with just those words. Thanks for sharing John. That's a really a neat idea, even if the results don't seem particularly relevant for some tests I tried. I was curious how it does the matching of page text against the profiles. I see from the description at http://vivosearchlight.org that EleasticSearch is being used instead of Solr. Any chance Miles Worthington (ok I googled) would be willing to share the source code on his github account [1], or elsewhere? //Ed [1] https://github.com/milesworthington
