SOLR search enables autosuggest, error correction, synonym consideration, image search and search within PDF files. It improves the user experience, creates real added value for visitors and thereby implicitly improves search engine rankings.
SOLR is a stand-alone search engine software developed as open source by Apache Software Foundation. With SOLR, website operators can offer visitors a powerful search function. This includes, in addition to pure full text search, autocomplete, right spelling corrections, consideration of synonyms and root word variants, rich content support (search in PDFs, Word files, etc.), text highlighting as well as faceting. Furthermore, certain areas of the website can be weighted differently. For example, the most recent news articles can be given the most weight.
To use SOLR you need:
More information about SOLR at: http://lucene.apache.org/solr/
For small websites TYPO3 offers an integrated search function: indexed_search. This is easy to set up and works reliably. A page is indexed by the extension the first time it is called up in the browser. It doesn't matter if it's normal pages, single views of extensions like news or even the output of own extensions. indexed_search simply records all words and terms output on the page into corresponding tables of the database. If a visitor then searches for a term, all pages containing this term are displayed in the search result.
For small websites, this is a good way to implement a search that finds everything displayed on the website without much effort.
For more extensive websites, however, indexed_search slows down the system quite quickly. The tables can quickly become very large when using event plugins, news, etc. intensively.
In addition, there is a lack of functions that make the search more user-friendly and the search results more meaningful: searching by criteria (faceted search), support for misspelled search terms (Do you mean ...), defining stop words (and, is...) that are not taken into account in the search or weighting the search results (e.g. according to topicality or type of entry) are not possible.
Basically, the standard search of TYPO3 - since it is a pure full text search - is not able to search for multiple terms. A search query "TYPO3 agency" would therefore only find the places where the text is exactly in the order. Pages with the terms "agencies for TYPO3", "TYPO3 internet agencies" or "web agency for TYPO3", for example, would not be considered.
For the setup of the TYPO3 extension solr a SOLR server is required. Many providers offer SOLR servers - similar to normal web hosting - for a flat monthly fee.
The extension is installed via the extension manager in TYPO3 and the connection with the Solr core is set up. The various modules for indexing are then set up in Typoscript.
For multilingual websites, a separate core is included for each language, so that the configuration of Solr is adapted to the respective language.
For the indexing of the texts on the pages and the news extension, the support is set up in the Typoscript templates of the extension. In the HTML template the markers
TYPO3SEARCH_begin and TYPO3SEARCH_end
must be included in the HTML template and the indexing on the web page must be activated in Typoscript.
config.index_enable = 1
For all other extensions the indexing is configured via Typoscript. The search goes directly via the database tables. The weighting of the search results between pages and extensions can also be changed, but also the weighting of different fields such as title, abstract and text and even the weighting of individual tags can be adjusted.
Indexing is done continuously via a cronjob. When changes are made to the data records, the corresponding data record is automatically placed anew in the index queue, so that the search results always remain up-to-date. The newest data records are indexed first, so that the search delivers good results right from the start.
The search mask offers various supports in addition to the search field. For example, the faceted search can be used to select that only search results from the news or events are displayed.
Already when entering the search term, search terms can be suggested with Ajax, so that the user can select the best term for him. Even with the search results, a suggestion for an alternative search can help the user to improve his search query.
The most recent or particularly frequently entered search terms can be displayed so that the user can obtain the desired information with one click.
In addition to the title of the page, the search results also show with a graphic how relevant the search result is for the searched term. In the teaser text, the search term is highlighted, and phrases that are not relevant are excluded. This makes it easier for the visitor to see whether the page really contains the search term in the desired context.
The indexing of pdf files (or other file types) on the server is also possible with SOLR. For this purpose, the extensions tika and solrfal are included and configured. Of course, it is important that the metadata (title and description) of the pdf files are set correctly so that the files are displayed reasonably in the search results. The indexing of the files also runs via a cronjob.
The use of the extension solrfal requires the payment of a one-time development fee (EB). The EB is generally applicable to a current LTS version (and all versions below). When upgrading to a new TYPO3 version, the purchase of a new EB is usually required again.
The EB amounts - depending on your age - between 800, - € and 1,200, - €.
However, the investment is worthwhile, especially since the extension brings other advantages in addition to file indexing.
With solrfal
From our own website, we have installed SOLR on the following TYPO3 customer sites, among others:
SOLR can also be configured to search via custom tables and then display a custom results list. We have implemented this logic for a customer's course and appointment search (www.vfb-weiterbildung.de) . In this case, a search must be performed using a course table that is linked to a table for appointments. In the search result, certain fields of the two tables (e.g. course name, price, start, location, etc.) should then be displayed:
The following table shows an example cost calculation for the integration of SOLR search on a TYOP3 website.
| Our service | Costs* |
|---|---|
|
Integration of SOLR search server
|
Effort: approx. 16 Hrs. Non-profit organisations: Company: |
* all prices are exclusive of VAT.
| Integration of SOLR search server |
|---|
|
Costs* Effort: approx. 16 Hrs. Non-profit organisations: Company: |
* all prices are exclusive of VAT.
To use the SOLR search, you need your own (virtual) SOLR server. The software for this is free of charge (open source). However, in-depth technical knowledge is required for installation, configuration and operation. We therefore offer our customers the use of our powerful SOLR servers.
The following costs are incurred (a core is required for each language (German, English, etc.)):
| Our service | 1. core | each additional core |
|---|---|---|
| Provision SOLR server (32MB per core) | 9,00 €/mtl. | 5,00 €/mtl. |
Version matrix (which versions are compatible with each other): https://docs.typo3.org/p/apache-solr-for-typo3/solr/11.0/en-us/Appendix/VersionMatrix.html
