yuuvis® Momentum works with Elasticsearch as search engine, combined with plug-ins.

Introduction

yuuvis® Momentum requires a high-performant search engine, such as Elasticsearch in combination with the IntraFind Linguistics Plug-in and the ICU Analysis Plug-in.
>> Supported Elasticsearch Versions

The search behavior in yuuvis® Momentum is affected by the configuration of Elasticsearch and the corresponding plug-ins. This article collects some important and interesting notes on this topic.

General Notes regarding Search Behavior

Paging

Paging is used to retrieve a section of a long result list. Those sections are considered as pages. The number of results per page as well as the first result to be included can be specified in each search request in yuuvis® Momentum.
>> Search Query Language

In order to ensure high performance, yuuvis® Momentum does not support the usage of a cursor in search requests. This can have an effect on the paging behavior in following situation:

• the total number of results is larger than the number of displayed results per page AND
• some objects matching the query condition are created/updated/deleted before the user retrieves another page of results.

Since the retrieval of another page means a new search request, the total result list is different. Thus, the requested page might contain unexpected results, e.g., results that were already displayed in on the previous page.

Applying a new Configuration

If you changed the index configuration, all objects that are imported afterwards will be indexed as defined by the new configuration. All objects that are already in the system remain unchanged and can be searched as before the configuration update. If you need all objects indexed with refer to the new configuration, you need a reindexing.
>> Elasticsearch Reindex

Notes on IntraFind Linguistics Plug-in

The plug-in offers semantic and linguistic text analysis features as well as powerful query operators.
>> IntraFind Linguistics Plug-in
>> Semantic-linguistic Indexing
>> Stemming and Lemmatization

Supported Languages

With the standard installation of yuuvis® Momentum, the plug-in supports a selection of languages as shown in the table below. ONE of them at time can be used with the standard license. Per default, German is activated. An alternative language can be activated instead of German. To activate more than one language, a license extension is required.

Per default, the core system is configured to support German and English language-specific analysis. Alternative languages to be used instead can be configured in the application-es.yml configuration file. If you want to apply configuration changes to a system already in operation, please note the general remarks above.

Language	ISO 639-1	Supported in default analyzer configuration	Intrafind license in standard installations
			Activated in standard license	Configurable alternative languages
English	en		-
French	fr	-	-
German	de
Italian	it	-	-
Spanish	es	-	-

A variety of further languages can be supported by the plug-in. Please contact IntraFind to get detailed information. You need to obtain a corresponding license on your own responsibility.

Language-specific Examples

Japanese

Standard Analysis

If the Elasticsearch standard analyzer is configured to support language-specific full-text analysis for Japanese, Tokenization and Keyword Analysis are supported.

E.g., "春夏秋冬" will be analyzed as [春, 夏, 秋, 冬], and then can be searched by 春 or 夏 or 秋 or 冬.

Most punctuation marks are removed. Upper case is changed into lower case. For "夏見、おはよう", e.g., 夏見 will be recognized as a searchable keyword.

With Japanese IntraFind Linguistics Plug-in License

If a license enabling the plug-in's functionality for Japanese is additionally activated, Japanese words can be extracted from phrases but not all.

E.g., for "寿司がおいしいね", 寿司 will be recognized as a searchable keyword.