Hey all,
One thing I have always struggled with on MODX is a good, proper useful search function. I have used simplesearch, advanced search, msearch… all of them generally leave me lacking as far as good end user searches go.
It used to be easier when google allowed for embedded searching.
Does anyone have any tips or guides for implementing a really good site search? I have good content - everything is SEO friendly and contextually loaded properly.
Thanks!
The problem with the site searches is that without a proper search engine in the back end, everything will just be varying forms of increasingly complex MySQL queries.
Apache Solr and OpenSearch are search engine servers that do much more than basic queries. The search at extras.modx.com is on OpenSearch.
That said, I’m not sure if you ever worked with the field potency settings in Simple Search. That can dramatically improve the results for Resources.
I have played a little bit with those setting and it definitely helped - prioritizing fields and content weights are big help.
One thing - and I completely understand this is outside the purview and expectations for MODX - is document indexing support - i.e. PDF primarily.
I know SOLR (will explore OpenSearch) supports such indexing, but I think I’ve attempted to implement this tool probably four or five times before giving up with some measure of frustration.
Granted, it’s been a couple of years. At a minimum, thanks for the tip about OpenSearch - I’ll check it out
OpenSearch is the maintained open source version of Elastic Search. It’s pretty high performance, but it will still require a server to run on.
By the way, Google Programmable Search does still work and provided you’re on a paid Google Development plan, there are no ads served.
Does “but it will still require a server to run on” mean it will run on the MODXcloud?
Asking because I have several long Index PDFs that need searching:
I converted most from PDFs to HTML only to spend countless hours getting rid of all the miscellaneous HTML to build ULs.
Now the client wants to go back to the Original PDFs and wants them to be Searchable.
I’ve toyed with the idea of using Tags but with thousands of entries overall and the client building the PDFs in Pages on an ancient iMac (!!), I skipped that in 30 seconds. Better to have something Search and Index the PDF contents by name rather than scroll through the PDFs.