{"id":518,"date":"2010-07-15T09:49:52","date_gmt":"2010-07-15T09:49:52","guid":{"rendered":"https:\/\/www.shapeways.com\/blog\/archives\/518-major-improvements-to-product-and-site-search-on-shapeways.html"},"modified":"2014-07-03T16:57:02","modified_gmt":"2014-07-03T16:57:02","slug":"major-improvements-to-product-and-site-search-on-shapeways","status":"publish","type":"post","link":"https:\/\/www.shapeways.com\/blog\/major-improvements-to-product-and-site-search-on-shapeways","title":{"rendered":"Major improvements to product and site search on Shapeways"},"content":{"rendered":"

\nThe last couple of weeks we have updated both our site-wide search and
\nour product search. Our previous search options were very limited in
\nfunctionality. With the growth of our content — both in articles as in
\nproducts — it was getting very hard to find things on shapeways.com. I
\nwould like to take this opportunity to tell you a little bit on what we
\nhave developed, how you can use it and how to improve the ability to
\nfind your products by our visitors.\n<\/p>\n<\/p>\n

<\/p>\n

The first update was our site-wide search<\/a>.
\nWe have opted to integrate Google site search for this feature. Google
\nsite search offers the most scalable solution and with the use of
\nGoogle’s page ranking algorithm the results were the best of all
\nsolutions we evaluated. We use the result set of Google and present
\nthis to the user. We do not influence the result set in any way. We do
\nenrich the result set for shops, users and products with a thumbnail.
\nAt this moment the total search set contains 175.000 pages and is
\ngrowing every day.<\/div>\n
 <\/div>\n
The second update was a complete new implementation of product
\nsearch<\/a>.
\nWe opted for Apache Solr<\/a>. This
\nsolution offers excellent search results and scales very well. Moreover
\nit gives us the ability to implement faceted search. With faceted
\nsearch we can determine relevant categories, tags, shops, etc based on
\nthe search terms and help the user to drill down to the most relevant
\ncontent. But this is something for later. We now focus on plain search
\nonly.<\/div>\n
We index the following attributes of products:<\/div>\n