Site wide search Seek and ye shall find.
With increasing content, it is usually not easier to present the desired directly to the visitor. To make a fast navigation to the "wanted" possible, I decided to implement a full text search with autocompletion based on Apache Solr also on this website.
Enterprise Search with Solr
Here you see my implementation of the search solution Apache Solr for TYPO3 CMS
The search results were specially tailored to the needs of the TYPONiels presence and enhanced with features such as boosting, typo correction and auto-complete.
Frequently asked questions & answers
The answer to the following questions may also interest you
The search function on typoniels.de differs somewhat from the classic search functions in CMS-driven websites, because in my website the search is not performed directly in the database (database index), but the content is stored (indexed) in an external Apache Solr search server and can thus be delivered even faster and more efficiently within a few milliseconds. For this to work, the content of the page must be transferred from the CMS to the database of the search server in structured form using a specially developed script and continuously kept synchronized there.
Once this workflow is set up, you get an even more powerful search, which then has no problems even with very large data volumes (10,000 documents and more). This performance makes it possible to offer functions such as auto-completion, typo correction and synonym recognition.
Yes, the search function on TYPONiels.de is built in such a way that content from other websites is also integrated and displayed. For example, content from the technology radar at radar.niels-langlotz.com is also indexed.
Update 2021: In the course of switching to Elasticsearch as a headless content repository, my goal is to merge my web infrastructure even further and make content usable for the editor across the board.
The costs for an individual search function are difficult to narrow down across the board and always depend on the complexity and handling or uniformity of the expected database. For very branched content structures, the implementation is more costly than for pages with one or two dominant content types.
In general, the basis of the search function, the software Apache Solr is an open source product and is therefore already available free of charge. Costs arise here only for the setup, connection to your system and the desired customizations and individualization. As a rule, you can calculate with 2 to 5 person-days.
Yes, external content can be brought into the search index very easily due to the dedicated search server. Here, in the end, you only have to decide whether a separation of the sources into different indexes (data stores) must take place or whether a mixture may take place. With Apache Nutch, it is also possible to crawl the source code of existing websites directly without having to index them via the REST API.
With the content elevation function, it is also possible to enhance external content, e.g. for marketing campaigns, and thus, among other things, to display offers from your own store prominently in the search of the corporate website.
I would be happy to advise you on this topic.
