Full text search

Visitors can use efficient search functions to find the information they need on the internet and intranet.
This means that any number of complex search queries can be made on databases. Website visitors simply enter the relevant values for the search in freely definable input masks and the corresponding search results are displayed automatically.
Various search systems and functions are available for searching websites, depending on the required scope of services (see also price list).

• Full-text search basis
  Professional search function in the website with indexing of content (without personalized content)
  
• Search statistics
  Tag clouds for statistics and websites, time history analysis to optimize meta data and navigation
• PDF search
  Search also captures the content of PDF files whose content can be read.
  
• Microsoft Word-^™, Excel-^™, Powerpoint-^™ search
  Search also captures Office and OpenOffice file types (docx, xlsx, pptx, odt, ods, odp).
• Personalized search
  Search also captures personalized content
• AI support full-text search
  Search supplemented by vectors for AI search (embedding model)

The basic module allows you to search the markup files linked via the presence for search terms. The search accesses the active search index. When the search index is created, the search goes over the pages like a spider (robot / bot) and thus captures the page content generated when it is called up. By default, the search already captures information from the document extension of a checked-in binary file (title, description and keywords).
This enables visitors to a website to search for specific content and words.

A similarity search can already be used in the standard system (suggestion for similar terms). This is based on the Metaphone algorithm. If a specific search term does not allow any corresponding hits, the search function suggests possible similar terms as well as slightly differently spelled terms as a new search term for which corresponding search results exist. For example, if no results are found for the search term "content", the system suggests the alternative: "Did you mean "content"?", provided the word "content" appears on the website.
You can also use Soundex as an alternative to Metaphone. To do this, manually set the corresponding node in the relevant search configuration in the XML: /wd:robot/similarityAlgorithm/@value = "soundex" and then run the search index again.

As an extension to the basic module, search statistics offers the option of using tag clouds for statistical purposes and for output in the presence. It can also be used for time history analysis to optimize metadata and navigation.

If the content of binary documents also needs to be captured and searched, the PDF search and Microsoft Word^™, Excel^™, PowerPoint^™ search add-on modules provide two functions that offer this option.

Often entire directory structures with e.g. Word^™ and/or PDF documents have to be published on the Internet or intranet. Finding certain documents, especially large document inventories, particularly in existing directory structures, is of great importance.
With these extensions, Weblication^® has a powerful full-text search that not only searches HTML/XHTML documents, but also the binary files mentioned in full text.

Weblication^® indexes all markup, Word^™, Excel^™, PowerPoint^™ and PDF files, for which we recommend the use of external applications. These documents can also be found quickly and easily via a full-text search on the website and intranet.

Please note that only unencrypted documents can be searched and indexed!

In protected areas that are implemented via the Personalization add-on module, the Personalized Search add-on module can also be used to search through the personalized content and display it to users who are authorized to view it.

When using the AI search (additional module / activation in the search settings (AI)), vectors are created parallel to the search index during search indexing. The content found for the search (without personalized pages) is filtered so that, for example, no more images are included. These are then vectorized using an embedding model.

These vectors are used for the search results depending on the weightings and threshold values set in the search results mask, which is usually located under /CONTENT PROJECT/search/index.php.

In this way, the search results can be enhanced by the AI and, among other things, bring up results that the normal full-text search does not display.

The AI search can also be used to ask questions such as: "What is CSS-Only for?"