Chapter 45. Zend_Search_Lucene

Table of Contents

45.1. Overview
45.1.1. Introduction
45.1.2. Document and Field Objects
45.1.3. Understanding Field Types
45.1.4. HTML documents
45.1.5. Word 2007 documents
45.1.6. Powerpoint 2007 documents
45.1.7. Excel 2007 documents
45.2. Building Indexes
45.2.1. Creating a New Index
45.2.2. Updating Index
45.2.3. Updating Documents
45.2.4. Retrieving Index Size
45.2.5. Index optimization
45.2.5.1. MaxBufferedDocs auto-optimization option
45.2.5.2. MaxMergeDocs auto-optimization option
45.2.5.3. MergeFactor auto-optimization option
45.2.6. Permissions
45.2.7. Limitations
45.2.7.1. Index size
45.2.7.2. Supported Filesystems
45.3. Searching an Index
45.3.1. Building Queries
45.3.1.1. Query Parsing
45.3.2. Search Results
45.3.3. Limiting the Result Set
45.3.4. Results Scoring
45.3.5. Search Result Sorting
45.3.6. Search Results Highlighting
45.4. Query Language
45.4.1. Terms
45.4.2. Fields
45.4.3. Wildcards
45.4.4. Term Modifiers
45.4.5. Range Searches
45.4.6. Fuzzy Searches
45.4.7. Matched terms limitation
45.4.8. Proximity Searches
45.4.9. Boosting a Term
45.4.10. Boolean Operators
45.4.10.1. AND
45.4.10.2. OR
45.4.10.3. NOT
45.4.10.4. &&, ||, and ! operators
45.4.10.5. +
45.4.10.6. -
45.4.10.7. No Operator
45.4.11. Grouping
45.4.12. Field Grouping
45.4.13. Escaping Special Characters
45.5. Query Construction API
45.5.1. Query Parser Exceptions
45.5.2. Term Query
45.5.3. Multi-Term Query
45.5.4. Boolean Query
45.5.5. Wildcard Query
45.5.6. Fuzzy Query
45.5.7. Phrase Query
45.5.8. Range Query
45.6. Character Set
45.6.1. UTF-8 and single-byte character set support
45.6.2. Default text analyzer
45.6.3. UTF-8 compatible text analyzers
45.7. Extensibility
45.7.1. Text Analysis
45.7.2. Tokens Filtering
45.7.3. Scoring Algorithms
45.7.4. Storage Containers
45.8. Interoperating with Java Lucene
45.8.1. File Formats
45.8.2. Index Directory
45.8.3. Java Source Code
45.9. Advanced
45.9.1. Starting from 1.6, handling index format transformations
45.9.2. Using the index as static property
45.10. Best Practices
45.10.1. Field names
45.10.2. Indexing performance
45.10.3. Index during Shut Down
45.10.4. Retrieving documents by unique id
45.10.5. Memory Usage
45.10.6. Encoding
45.10.7. Index maintenance

45.1. Overview

45.1.1. Introduction

Zend_Search_Lucene is a general purpose text search engine written entirely in PHP 5. Since it stores its index on the filesystem and does not require a database server, it can add search capabilities to almost any PHP-driven website. Zend_Search_Lucene supports the following features:

  • Ranked searching - best results returned first

  • Many powerful query types: phrase queries, boolean queries, wildcard queries, proximity queries, range queries and many others.

  • Search by specific field (e.g., title, author, contents)

Zend_Search_Lucene was derived from the Apache Lucene project. The currently (starting from ZF 1.6) supported Lucene index format versions are 1.4 - 2.3. For more information on Lucene, visit http://lucene.apache.org/java/docs/.

[Note]

Previous Zend_Search_Lucene implementations support the Lucene 1.4 (1.9) - 2.1 index formats.

Starting from ZF 1.5 any index created using pre-2.1 index format is automatically upgraded to Lucene 2.1 format after the Zend_Search_Lucene update and will not be compatible with Zend_Search_Lucene implementations included into ZF 1.0.x.

45.1.2. Document and Field Objects

Zend_Search_Lucene operates with documents as atomic objects for indexing. A document is divided into named fields, and fields have content that can be searched.

A document is represented by the Zend_Search_Lucene_Document class, and this objects of this class contain instances of Zend_Search_Lucene_Field that represent the fields on the document.

It is important to note that any information can be added to the index. Application-specific information or metadata can be stored in the document fields, and later retrieved with the document during search.

It is the responsibility of your application to control the indexer. This means that data can be indexed from any source that is accessible by your application. For example, this could be the filesystem, a database, an HTML form, etc.

Zend_Search_Lucene_Field class provides several static methods to create fields with different characteristics:

$doc = new Zend_Search_Lucene_Document();

// Field is not tokenized, but is indexed and stored within the index.
// Stored fields can be retrived from the index.
$doc->addField(Zend_Search_Lucene_Field::Keyword('doctype',
                                                 'autogenerated'));

// Field is not tokenized nor indexed, but is stored in the index.
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('created',
                                                   time()));

// Binary String valued Field that is not tokenized nor indexed,
// but is stored in the index.
$doc->addField(Zend_Search_Lucene_Field::Binary('icon',
                                                $iconData));

// Field is tokenized and indexed, and is stored in the index.
$doc->addField(Zend_Search_Lucene_Field::Text('annotation',
                                              'Document annotation text'));

// Field is tokenized and indexed, but is not stored in the index.
$doc->addField(Zend_Search_Lucene_Field::UnStored('contents',
                                                  'My document content'));

Each of these methods (excluding the Zend_Search_Lucene_Field::Binary() method) has an optional $encoding parameter for specifying input data encoding.

Encoding may differ for different documents as well as for different fields within one document:

$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Text('title',
                                              $title,
                                              'iso-8859-1'));
$doc->addField(Zend_Search_Lucene_Field::UnStored('contents',
                                                  $contents,
                                                  'utf-8'));

If encoding parameter is omitted, then the current locale is used at processing time. For example:

setlocale(LC_ALL, 'de_DE.iso-8859-1');
...
$doc->addField(Zend_Search_Lucene_Field::UnStored('contents', $contents));

Fields are always stored and returned from the index in UTF-8 encoding. Any required conversion to UTF-8 happens automatically.

Text analyzers (see below) may also convert text to some other encodings. Actually, the default analyzer converts text to 'ASCII//TRANSLIT' encoding. Be careful, however; this translation may depend on current locale.

Fields' names are defined at your discretion in the addField() method.

Java Lucene uses the 'contents' field as a default field to search. Zend_Search_Lucene searches through all fields by default, but the behavior is configurable. See the "Default search field" chapter for details.

45.1.3. Understanding Field Types

  • Keyword fields are stored and indexed, meaning that they can be searched as well as displayed in search results. They are not split up into separate words by tokenization. Enumerated database fields usually translate well to Keyword fields in Zend_Search_Lucene.

  • UnIndexed fields are not searchable, but they are returned with search hits. Database timestamps, primary keys, file system paths, and other external identifiers are good candidates for UnIndexed fields.

  • Binary fields are not tokenized or indexed, but are stored for retrieval with search hits. They can be used to store any data encoded as a binary string, such as an image icon.

  • Text fields are stored, indexed, and tokenized. Text fields are appropriate for storing information like subjects and titles that need to be searchable as well as returned with search results.

  • UnStored fields are tokenized and indexed, but not stored in the index. Large amounts of text are best indexed using this type of field. Storing data creates a larger index on disk, so if you need to search but not redisplay the data, use an UnStored field. UnStored fields are practical when using a Zend_Search_Lucene index in combination with a relational database. You can index large data fields with UnStored fields for searching, and retrieve them from your relational database by using a separate field as an identifier.

    Table 45.1. Zend_Search_Lucene_Field Types

    Field Type Stored Indexed Tokenized Binary
    Keyword Yes Yes No No
    UnIndexed Yes No No No
    Binary Yes No No Yes
    Text Yes Yes Yes No
    UnStored No Yes Yes No

45.1.4. HTML documents

Zend_Search_Lucene offers a HTML parsing feature. Documents can be created directly from a HTML file or string:

$doc = Zend_Search_Lucene_Document_Html::loadHTMLFile($filename);
$index->addDocument($doc);
...
$doc = Zend_Search_Lucene_Document_Html::loadHTML($htmlString);
$index->addDocument($doc);

Zend_Search_Lucene_Document_Html class uses the DOMDocument::loadHTML() and DOMDocument::loadHTMLFile() methods to parse the source HTML, so it doesn't need HTML to be well formed or to be XHTML. On the other hand, it's sensitive to the encoding specified by the "meta http-equiv" header tag.

Zend_Search_Lucene_Document_Html class recognizes document title, body and document header meta tags.

The 'title' field is actually the /html/head/title value. It's stored within the index, tokenized and available for search.

The 'body' field is the actual body content of the HTML file or string. It doesn't include scripts, comments or attributes.

The loadHTML() and loadHTMLFile() methods of Zend_Search_Lucene_Document_Html class also have second optional argument. If it's set to true, then body content is also stored within index and can be retrieved from the index. By default, the body is tokenized and indexed, but not stored.

The third parameter of loadHTML() and loadHTMLFile() methods optionally specifies source HTML document encoding. It's used if encoding is not specified using Content-type HTTP-EQUIV meta tag.

Other document header meta tags produce additional document fields. The field 'name' is taken from 'name' attribute, and the 'content' attribute populates the field 'value'. Both are tokenized, indexed and stored, so documents may be searched by their meta tags (for example, by keywords).

Parsed documents may be augmented by the programmer with any other field:

$doc = Zend_Search_Lucene_Document_Html::loadHTML($htmlString);
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('created',
                                                   time()));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('updated',
                                                   time()));
$doc->addField(Zend_Search_Lucene_Field::Text('annotation',
                                              'Document annotation text'));
$index->addDocument($doc);

Document links are not included in the generated document, but may be retrieved with the Zend_Search_Lucene_Document_Html::getLinks() and Zend_Search_Lucene_Document_Html::getHeaderLinks() methods:

$doc = Zend_Search_Lucene_Document_Html::loadHTML($htmlString);
$linksArray = $doc->getLinks();
$headerLinksArray = $doc->getHeaderLinks();

Starting from ZF 1.6 it's also possible to exclude links with rel attribute set to 'nofollow'. Use Zend_Search_Lucene_Document_Html::setExcludeNoFollowLinks($true) to turn on this option.

Zend_Search_Lucene_Document_Html::getExcludeNoFollowLinks() method returns current state of "Exclude nofollow links" flag.

45.1.5. Word 2007 documents

Zend_Search_Lucene offers a Word 2007 parsing feature. Documents can be created directly from a Word 2007 file:

$doc = Zend_Search_Lucene_Document_Docx::loadDocxFile($filename);
$index->addDocument($doc);

Zend_Search_Lucene_Document_Docx class uses the ZipArchive class and simplexml methods to parse the source document. If the ZipArchive class (from module php_zip) is not available, the Zend_Search_Lucene_Document_Docx will also not be available for use with Zend Framework.

Zend_Search_Lucene_Document_Docx class recognizes document meta data and document text. Meta data consists, depending on document contents, of filename, title, subject, creator, keywords, description, lastModifiedBy, revision, modified, created.

The 'filename' field is the actual Word 2007 file name.

The 'title' field is the actual document title.

The 'subject' field is the actual document subject.

The 'creator' field is the actual document creator.

The 'keywords' field contains the actual document keywords.

The 'description' field is the actual document description.

The 'lastModifiedBy' field is the username who has last modified the actual document.

The 'revision' field is the actual document revision number.

The 'modified' field is the actual document last modified date / time.

The 'created' field is the actual document creation date / time.

The 'body' field is the actual body content of the Word 2007 document. It only includes normal text, comments and revisions are not included.

The loadDocxFile() methods of Zend_Search_Lucene_Document_Docx class also have second optional argument. If it's set to true, then body content is also stored within index and can be retrieved from the index. By default, the body is tokenized and indexed, but not stored.

Parsed documents may be augmented by the programmer with any other field:

$doc = Zend_Search_Lucene_Document_Docx::loadDocxFile($filename);
$doc->addField(Zend_Search_Lucene_Field::UnIndexed(
    'indexTime',
    time())
);
$doc->addField(Zend_Search_Lucene_Field::Text(
    'annotation',
    'Document annotation text')
);
$index->addDocument($doc);

45.1.6. Powerpoint 2007 documents

Zend_Search_Lucene offers a Powerpoint 2007 parsing feature. Documents can be created directly from a Powerpoint 2007 file:

$doc = Zend_Search_Lucene_Document_Pptx::loadPptxFile($filename);
$index->addDocument($doc);

Zend_Search_Lucene_Document_Pptx class uses the ZipArchive class and simplexml methods to parse the source document. If the ZipArchive class (from module php_zip) is not available, the Zend_Search_Lucene_Document_Pptx will also not be available for use with Zend Framework.

Zend_Search_Lucene_Document_Pptx class recognizes document meta data and document text. Meta data consists, depending on document contents, of filename, title, subject, creator, keywords, description, lastModifiedBy, revision, modified, created.

The 'filename' field is the actual Powerpoint 2007 file name.

The 'title' field is the actual document title.

The 'subject' field is the actual document subject.

The 'creator' field is the actual document creator.

The 'keywords' field contains the actual document keywords.

The 'description' field is the actual document description.

The 'lastModifiedBy' field is the username who has last modified the actual document.

The 'revision' field is the actual document revision number.

The 'modified' field is the actual document last modified date / time.

The 'created' field is the actual document creation date / time.

The 'body' field is the actual content of all slides and slide notes in the Powerpoint 2007 document.

The loadPptxFile() methods of Zend_Search_Lucene_Document_Pptx class also have second optional argument. If it's set to true, then body content is also stored within index and can be retrieved from the index. By default, the body is tokenized and indexed, but not stored.

Parsed documents may be augmented by the programmer with any other field:

$doc = Zend_Search_Lucene_Document_Pptx::loadPptxFile($filename);
$doc->addField(Zend_Search_Lucene_Field::UnIndexed(
    'indexTime',
    time()));
$doc->addField(Zend_Search_Lucene_Field::Text(
    'annotation',
    'Document annotation text'));
$index->addDocument($doc);

45.1.7. Excel 2007 documents

Zend_Search_Lucene offers a Excel 2007 parsing feature. Documents can be created directly from a Excel 2007 file:

$doc = Zend_Search_Lucene_Document_Xlsx::loadXlsxFile($filename);
$index->addDocument($doc);

Zend_Search_Lucene_Document_Xlsx class uses the ZipArchive class and simplexml methods to parse the source document. If the ZipArchive class (from module php_zip) is not available, the Zend_Search_Lucene_Document_Xlsx will also not be available for use with Zend Framework.

Zend_Search_Lucene_Document_Xlsx class recognizes document meta data and document text. Meta data consists, depending on document contents, of filename, title, subject, creator, keywords, description, lastModifiedBy, revision, modified, created.

The 'filename' field is the actual Excel 2007 file name.

The 'title' field is the actual document title.

The 'subject' field is the actual document subject.

The 'creator' field is the actual document creator.

The 'keywords' field contains the actual document keywords.

The 'description' field is the actual document description.

The 'lastModifiedBy' field is the username who has last modified the actual document.

The 'revision' field is the actual document revision number.

The 'modified' field is the actual document last modified date / time.

The 'created' field is the actual document creation date / time.

The 'body' field is the actual content of all cells in all worksheets of the Excel 2007 document.

The loadXlsxFile() methods of Zend_Search_Lucene_Document_Xlsx class also have second optional argument. If it's set to true, then body content is also stored within index and can be retrieved from the index. By default, the body is tokenized and indexed, but not stored.

Parsed documents may be augmented by the programmer with any other field:

$doc = Zend_Search_Lucene_Document_Xlsx::loadXlsxFile($filename);
$doc->addField(Zend_Search_Lucene_Field::UnIndexed(
    'indexTime',
    time()));
$doc->addField(Zend_Search_Lucene_Field::Text(
    'annotation',
    'Document annotation text'));
$index->addDocument($doc);