Full-Text

This article is part of the XQuery Portal. It summarizes the full-text features of BaseX.

Full-text retrieval is an essential query feature for working with XML documents, and BaseX was the first query processor that fully supported the W3C XQuery Full Text 1.0 Recommendation. This page lists some singularities and extensions of the BaseX implementation.

Query Evaluation

BaseX offers different evaluation strategies for XQFT queries, the choice of which depends on the input data and the existence of a full text index. The query compiler tries to optimize and speed up queries by applying a full text index structure whenever possible and useful. Three evaluation strategies are available: the standard sequential database scan, a full-text index based evaluation and a hybrid one, combining both strategies (see "XQuery Full Text implementation in BaseX"). Query optimization and selection of the most efficient evaluation strategy is done in a full-fledged automatic manner. The output of the query optimizer indicates which evaluation plan is chosen for a specific query. It can be inspected by activating verbose querying (Command: SET VERBOSE ON) or opening the Query Info in the GUI. The message

Applying full-text index

suggests that the full-text index is applied to speed up query evaluation. A second message

Removing path with no index results

indicates that the index does not yield any results for the specified term and is thus skipped. If index optimizations are missing, it sometimes helps to give the compiler a second chance and try different rewritings of the same query.

Options

The available full-text index can handle various combinations of the match options defined in the XQuery Full Text Recommendation. By default, most options are disabled. The GUI dialogs for creating new databases or displaying the database properties contain a tab for choosing between all available options. On the command-line, the SET command can be used to activate full-text indexing or creating a full-text index for existing databases:

• SET FTINDEX true; CREATE DB input.xml
• CREATE INDEX fulltext

The following indexing options are available:

• Language: see below for more details (SET LANGUAGE EN)
• Support Wildcards: a trie-based index can be applied to support wildcard searches (SET WILDCARDS true)
• Stemming: tokens are stemmed with the Porter Stemmer before being indexed (SET STEMMING true)
• Case Sensitive: tokens are indexed in case-sensitive mode (SET CASESENS true)
• Diacritics: diacritics are indexed as well (SET DIACRITICS true)
• TF/IDF Scoring: TF/IDF-based scoring values are calculated and stored in the index (SET SCORING 0/1/2; details see below)
• Stopword List: a stop word list can be defined to reduce the number of indexed tokens (SET STOPWORDS [filename])

Languages

The chosen language determines how the input text will be tokenized and stemmed. The basic code base and jar file of BaseX comes with built-in support for English and German. More languages are supported if the following libraries are found in the classpath:

• lucene-stemmers-3.4.0.jar: includes Snowball and Lucene stemmers and extends language support to the following languages: Arabic, Bulgarian, Catalan, Czech, Danish, Dutch, Finnish, French, Hindi, Hungarian, Italian, Latvian, Lithuanian, Norwegian, Portuguese, Romanian, Russian, Spanish, Swedish, Turkish.

• igo-0.4.3.jar: An additional article explains how Igo can be integrated, and how Japanese texts are tokenized and stemmed.

The JAR files can also be found in the zip and exe distribution files of BaseX.

The following two queries, which both return true, demonstrate that stemming depends on the selected language:

"Indexing" contains text "index" using stemming,
"häuser" contains text "haus" using stemming using language "de"

Scoring

The XQuery Full Text Recommendation allows for the usage of scoring models and values within queries, with scoring being completely implementation defined. BaseX offers an efficient internal scoring model which can be easily extended to different application scenarios. Additionally, BaseX allows to store scoring values within the full-text index structure (demanding additional time and memory). Three scoring types are currently available, which can be adjusted with the SCORING property (Default: SET SCORING 0):

• 0: This algorithm yields the best results for general-purpose use cases. It calculates the scoring value out of the length of a term and its frequency in a single text node. This algorithm is also applied if no index exists, or if the index cannot be applied in a query.
• 1: Standard TF/IDF algorithm, which treats document nodes as document units.
• 2: Each text node is treated as a document unit in the TF/IDF algorithm. This variant is an alternative for type 1, if the database contains large, few XML files.

Querying Using Thesaurus

BaseX supports full-text queries using thesauri, but it does not provide a default thesaurus. This is why query such as

'computers' contains text 'hardware'
  using thesaurus default

will return false. However, if the thesaurus is specified, then the result will be true

'computers' contains text 'hardware'
  using thesaurus at 'XQFTTS_1_0_4/TestSources/usability2.xml'

The format of the thesaurus files must be the same as the format of the thesauri provided by the XQuery and XPath Full Text 1.0 Test Suite. It is an XML with structure defined by an XSD Schema.

Fuzzy Querying

In addition to the official recommendation, BaseX supports fuzzy querying. The XQFT grammar was enhanced by the FTMatchOption using fuzzy to allow for approximate searches in full texts. By default, the standard full-text index already supports the efficient execution of fuzzy searches.

Document 'doc.xml':

<doc>
   <a>house</a>
   <a>hous</a>
   <a>haus</a>
</doc>

Command: CREATE DB doc.xml; CREATE INDEX fullext

Query:

//a[text() contains text 'house' using fuzzy]

Result:

<a>house</a>
<a>hous</a>

Fuzzy search is based on the Levenshtein distance. The maximum number of allowed errors is calculated by dividing the token length of a specified query term by 4, preserving a minimum of 1 errors. A static error distance can be set by adjusting the LSERROR property (default: SET LSERROR 0). The query above yields two results as there is no error between the query term "foo" and the text node "foo bar", and one error between "foo" and "foa bar".

Mixed Content

When working with so-called narrative XML documents, such as HTML, TEI, or DocBook documents, you typically have mixed content, i.e., elements containing a mix of text and markup, such as:

<p>This is only an illustrative <hi>example</hi>, not a <q>real</q> text.</p>

Since the logical flow of the text is not interrupted by the child elements, you will typically want to search across elements, so that the above paragraph would match a search for real text. For more examples, see XQuery and XPath Full Text 1.0 Use Cases.

To enable this kind of searches, whitespace chopping must be turned off by setting the option CHOP to OFF (default: SET CHOP ON). In the GUI, you find this option in Database → New… → Parsing → Chop Whitespaces.

A query such as //p[. contains text 'real text'] will then match the example paragraph above.

Note that the node structure is completely ignored by the full-text tokenizer: the contains text expression applies all full-text operations to the string value of its left operand. As a consequence, the ft:mark and ft:extract functions (see Full-Text Functions) will only yield useful results if they are applied to single text nodes, as the following example demonstrates:

(: Structure is ignored; no highlighting: :)
ft:mark(//p[. contains text 'real'])
(: Single text nodes are addressed: results will be highlighted: :)
ft:mark(//p[.//text() contains text 'real'])

Functions

Some additional Full-Text Functions have been added to BaseX to extend the official language recommendation with useful features, such as explicitly requesting the score value of an item, marking the hits of a full-text request, or directly accessing the full-text index with the default index options.