ninoskopac / phptikawrapper

Simple PHP Wrapper for Apache Tika

This is a simple PHP Wrapper for Apache Tika.

It allows the developer to retrieve text, metadata and language from complex documents.

Supported formats

It supports opendocument, office .doc and .docx, pdf, images, videos and a lot more !

See http://tika.apache.org/1.1/formats.html for details.

Install with composer

Add the package dependency enzim/tika-wrapper in your composer.json

    {
        "require": {
            "ninoskopac/php-tika-wrapper": "~1.0" 
        }   
    }

Install the new package with composer, and that's it!

php composer.phar install

For convenience, the package include the tika-app jar file, which is quite big (25MB). The download may take time!

See http://packagist.org for more details. (Don't forget to add require 'vendor/.composer/autoload.php'; in your autoloading php file).

Example installation/usage

See example/ (more docs to come soon) for an example:

    git clone [email protected]:pierroweb/PhpTikaWrapper.git
    cd PhpTikaWrapper
    
    cd example/with-composer
    curl -s http://getcomposer.org/installer | php
    php composer.phar install
    php usage.php

Usage

In your own project, assuming you have an opendocument test.odt in the current directory

    <?php
    use Enzim\Lib\TikaWrapper\TikaWrapper;

    $testFile = __DIR__."/test.odt";

    $plaintext = TikaWrapper::getText($testFile);
    $metadataArray = TikaWrapper::getMetaData($testFile);
    $language = TikaWrapper::getLanguage($testFile);

Available methods (they all take a string, the full path of the file, as argument)

• getText($file) returns a string containing the document in plain-text

• getTextMain($file) returns a string containing only the main text of the doc

• getXHTML($file) returns a string containing an XHTML (xml-valid) conversion of the document

• getHTML($file) returns a string containing an HTML conversion of the document

• getContentType($file) returns the content type of the document. Example outputs for opendocument, docx and pdf:

    application/vnd.oasis.opendocument.text
    application/vnd.openxmlformats-officedocument.wordprocessingml.document
    application/pdf
  

• getLanguage($file) returns the language of the documeent. Example output: en for english, fr for french, etc

• getMetaData($file) returns a PHP array with the metadata. Ex:

    Array  
    (
        [Character Count] => 41
        [Content-Length] => 8686
        [Content-Type] => application/vnd.oasis.opendocument.text
        [Creation-Date] => 2012-04-12T11:44:14
        [Edit-Time] => PT00H00M39S
        [Image-Count] => 0
        [Object-Count] => 0
        [Page-Count] => 1
        [Paragraph-Count] => 2
        [Table-Count] => 0
        [Word-Count] => 9
        [creator] => *******
        [date] => 2012-04-12T11:44:52
        [editing-cycles] => 1
        [generator] => OpenOffice.org/3.2$Linux OpenOffice.org_project/320m12$Build-9483
        [initial-creator] => Pierre B
        [nbCharacter] => 41
        [nbImg] => 0
        [nbObject] => 0
        [nbPage] => 1
        [nbPara] => 2
        [nbTab] => 0
        [nbWord] => 9
        [resourceName] => test.odt
        [xmpTPg:NPages] => 1
    )
  

TODO

• set a pretty print option (to use option -r for html/xhtml)
• allows the use of tika-server transparently to avoid loading the JVM on each request

Support for Tika-server

Supported by this lib: https://github.com/vaites/php-apache-tika

Credits

• http://tika.apache.org
• It uses the Symfony Process component http://symfony.com/doc/current/components/process.html