RTF reader: Unit tests and some improvements

ivanlanin · ivanlanin · commit 7a42802b4882 · 2014-05-29T18:02:49.000+07:00
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -4,7 +4,7 @@ This is the changelog between releases of PHPWord. Releases are listed in revers
 
 ## 0.11.0 - Not yet released
 
-This release marked the change of PHPWord license from LGPL 2.1 to LGPL 3. Three new elements were added: TextBox, ListItemRun, and Field. Relative and absolute positioning for images and textboxes were added. Writer classes were refactored into parts, elements, and styles. ODT and RTF features were enhanced. Ability to add elements to PHPWord object via HTML were implemeted.
+This release marked the change of PHPWord license from LGPL 2.1 to LGPL 3. Three new elements were added: TextBox, ListItemRun, and Field. Relative and absolute positioning for images and textboxes were added. Writer classes were refactored into parts, elements, and styles. ODT and RTF features were enhanced. Ability to add elements to PHPWord object via HTML were implemeted. RTF reader were initiated.
 
 ### Features
 
@@ -30,6 +30,7 @@ This release marked the change of PHPWord license from LGPL 2.1 to LGPL 3. Three
 - RTF Writer: Ability to write document properties - @ivanlanin
 - RTF Writer: Ability to write image - @ivanlanin
 - Element: New `Field` element - @basjan GH-251
+- RTF Reader: Basic RTF reader - @ivanlanin GH-72
 
 ### Bugfixes
 
diff --git a/docs/intro.rst b/docs/intro.rst
@@ -124,7 +124,7 @@ Readers
 +---------------------------+----------------------+--------+-------+-------+
 |                           | Custom               | ✓      |       |       |
 +---------------------------+----------------------+--------+-------+-------+
-| **Element Type**          | Text                 | ✓      | ✓     |       |
+| **Element Type**          | Text                 | ✓      | ✓     | ✓     |
 +---------------------------+----------------------+--------+-------+-------+
 |                           | Text Run             | ✓      |       |       |
 +---------------------------+----------------------+--------+-------+-------+
diff --git a/docs/src/documentation.md b/docs/src/documentation.md
@@ -114,7 +114,7 @@ Below are the supported features for each file formats.
 |-------------------------|--------------------|------|-----|-----|
 | **Document Properties** | Standard           | ✓    |     |     |
 |                         | Custom             | ✓    |     |     |
-| **Element Type**        | Text               | ✓    | ✓   |     |
+| **Element Type**        | Text               | ✓    | ✓   | ✓   |
 |                         | Text Run           | ✓    |     |     |
 |                         | Title              | ✓    | ✓   |     |
 |                         | Link               | ✓    |     |     |
diff --git a/samples/Sample_28_ReadRTF.php b/samples/Sample_28_ReadRTF.php
diff --git a/samples/resources/Sample_28_ReadRTF.rtf b/samples/resources/Sample_28_ReadRTF.rtf
diff --git a/src/PhpWord/Reader/RTF/Document.php b/src/PhpWord/Reader/RTF/Document.php
@@ -245,8 +245,8 @@ private function flush($isControl = false)
     private function flushControl($isControl = false)
     {
         if (preg_match("/^([A-Za-z]+)(-?[0-9]*) ?$/", $this->control, $match) === 1) {
-            list(, $control) = $match;
-            $this->parseControl($control);
+            list(, $control, $parameter) = $match;
+            $this->parseControl($control, $parameter);
         }
 
         if ($isControl === true) {
@@ -271,8 +271,7 @@ private function flushText()
 
             // Add text if it's not flagged as skipped
             if (!isset($this->flags['skipped'])) {
-                $textrun = $this->textrun->addText($this->text);
-                $this->flags['element'] = &$textrun;
+                $this->readText();
             }
 
             $this->text = '';
@@ -312,30 +311,36 @@ private function pushText($char)
      * @param string $control
      * @param string $parameter
      */
-    private function parseControl($control)
+    private function parseControl($control, $parameter)
     {
         $controls = array(
-            'par'       => array(self::PARA, 'paragraph',    true),
-            'b'         => array(self::STYL, 'bold',         true),
-            'i'         => array(self::STYL, 'italic',       true),
-            'u'         => array(self::STYL, 'underline',    true),
-            'fonttbl'   => array(self::SKIP, 'fonttbl',      null),
-            'colortbl'  => array(self::SKIP, 'colortbl',     null),
-            'info'      => array(self::SKIP, 'info',         null),
-            'generator' => array(self::SKIP, 'generator',    null),
-            'title'     => array(self::SKIP, 'title',        null),
-            'subject'   => array(self::SKIP, 'subject',      null),
-            'category'  => array(self::SKIP, 'category',     null),
-            'keywords'  => array(self::SKIP, 'keywords',     null),
-            'comment'   => array(self::SKIP, 'comment',      null),
-            'shppict'   => array(self::SKIP, 'pic',          null),
-            'fldinst'   => array(self::SKIP, 'link',         null),
+            'par'       => array(self::PARA,    'paragraph',    true),
+            'b'         => array(self::STYL,    'font',         'bold',         true),
+            'i'         => array(self::STYL,    'font',         'italic',       true),
+            'u'         => array(self::STYL,    'font',         'underline',    true),
+            'strike'    => array(self::STYL,    'font',         'strikethrough',true),
+            'fs'        => array(self::STYL,    'font',         'size',         $parameter),
+            'qc'        => array(self::STYL,    'paragraph',    'align',        'center'),
+            'sa'        => array(self::STYL,    'paragraph',    'spaceAfter',   $parameter),
+            'fonttbl'   => array(self::SKIP,    'fonttbl',      null),
+            'colortbl'  => array(self::SKIP,    'colortbl',     null),
+            'info'      => array(self::SKIP,    'info',         null),
+            'generator' => array(self::SKIP,    'generator',    null),
+            'title'     => array(self::SKIP,    'title',        null),
+            'subject'   => array(self::SKIP,    'subject',      null),
+            'category'  => array(self::SKIP,    'category',     null),
+            'keywords'  => array(self::SKIP,    'keywords',     null),
+            'comment'   => array(self::SKIP,    'comment',      null),
+            'shppict'   => array(self::SKIP,    'pic',          null),
+            'fldinst'   => array(self::SKIP,    'link',         null),
         );
 
         if (array_key_exists($control, $controls)) {
             list($function) = $controls[$control];
             if (method_exists($this, $function)) {
-                $this->$function($controls[$control]);
+                $directives = $controls[$control];
+                array_shift($directives); // remove the function variable; we won't need it
+                $this->$function($directives);
             }
         }
     }
@@ -347,7 +352,7 @@ private function parseControl($control)
      */
     private function readParagraph($directives)
     {
-        list(, $property, $value) = $directives;
+        list($property, $value) = $directives;
         $this->textrun = $this->section->addTextRun();
         $this->flags[$property] = $value;
     }
@@ -359,12 +364,8 @@ private function readParagraph($directives)
      */
     private function readStyle($directives)
     {
-        list(, $property, $value) = $directives;
-        $this->flags[$property] = $value;
-        if (isset($this->flags['element'])) {
-            $element = &$this->flags['element'];
-            $element->getFontStyle()->setStyleValue($property, $value);
-        }
+        list($style, $property, $value) = $directives;
+        $this->flags['styles'][$style][$property] = $value;
     }
 
     /**
@@ -374,8 +375,19 @@ private function readStyle($directives)
      */
     private function readSkip($directives)
     {
-        list(, $property) = $directives;
+        list($property) = $directives;
         $this->flags['property'] = $property;
         $this->flags['skipped'] = true;
     }
+
+    /**
+     * Read text
+     */
+    private function readText()
+    {
+        $text = $this->textrun->addText($this->text);
+        if (isset($this->flags['styles']['font'])) {
+            $text->getFontStyle()->setStyleByArray($this->flags['styles']['font']);
+        }
+    }
 }
diff --git a/tests/PhpWord/Tests/Reader/ODTextTest.php b/tests/PhpWord/Tests/Reader/ODTextTest.php
@@ -33,7 +33,7 @@ class ODTextTest extends \PHPUnit_Framework_TestCase
     public function testLoad()
     {
         $filename = __DIR__ . '/../_files/documents/reader.odt';
-        $object = IOFactory::load($filename, 'ODText');
-        $this->assertInstanceOf('PhpOffice\\PhpWord\\PhpWord', $object);
+        $phpWord = IOFactory::load($filename, 'ODText');
+        $this->assertInstanceOf('PhpOffice\\PhpWord\\PhpWord', $phpWord);
     }
 }
diff --git a/tests/PhpWord/Tests/Reader/RTFTest.php b/tests/PhpWord/Tests/Reader/RTFTest.php
@@ -0,0 +1,51 @@
+<?php
+/**
+ * This file is part of PHPWord - A pure PHP library for reading and writing
+ * word processing documents.
+ *
+ * PHPWord is free software distributed under the terms of the GNU Lesser
+ * General Public License version 3 as published by the Free Software Foundation.
+ *
+ * For the full copyright and license information, please read the LICENSE
+ * file that was distributed with this source code. For the full list of
+ * contributors, visit https://github.com/PHPOffice/PHPWord/contributors.
+ *
+ * @link        https://github.com/PHPOffice/PHPWord
+ * @copyright   2010-2014 PHPWord contributors
+ * @license     http://www.gnu.org/licenses/lgpl.txt LGPL version 3
+ */
+
+namespace PhpOffice\PhpWord\Tests\Reader;
+
+use PhpOffice\PhpWord\IOFactory;
+
+/**
+ * Test class for PhpOffice\PhpWord\Reader\RTF
+ *
+ * @coversDefaultClass \PhpOffice\PhpWord\Reader\RTF
+ * @runTestsInSeparateProcesses
+ */
+class RTFTest extends \PHPUnit_Framework_TestCase
+{
+    /**
+     * Test load
+     */
+    public function testLoad()
+    {
+        $filename = __DIR__ . '/../_files/documents/reader.rtf';
+        $phpWord = IOFactory::load($filename, 'RTF');
+        $this->assertInstanceOf('PhpOffice\\PhpWord\\PhpWord', $phpWord);
+    }
+
+    /**
+     * Test load exception
+     *
+     * @expectedException \Exception
+     * @expectedExceptionMessage Cannot read
+     */
+    public function testLoadException()
+    {
+        $filename = __DIR__ . '/../_files/documents/foo.rtf';
+        IOFactory::load($filename, 'RTF');
+    }
+}
diff --git a/tests/PhpWord/Tests/Reader/Word2007Test.php b/tests/PhpWord/Tests/Reader/Word2007Test.php
@@ -28,52 +28,33 @@
  */
 class Word2007Test extends \PHPUnit_Framework_TestCase
 {
-    /**
-     * Init
-     */
-    public function tearDown()
-    {
-    }
-
     /**
      * Test canRead() method
      */
     public function testCanRead()
     {
         $object = new Word2007();
-        $fqFilename = join(
-            DIRECTORY_SEPARATOR,
-            array(PHPWORD_TESTS_BASE_DIR, 'PhpWord', 'Tests', '_files', 'documents', 'reader.docx')
-        );
-        $this->assertTrue($object->canRead($fqFilename));
+        $filename = __DIR__ . '/../_files/documents/reader.docx';
+        $this->assertTrue($object->canRead($filename));
     }
 
     /**
      * Can read exception
-     *
-     * @expectedException \PhpOffice\PhpWord\Exception\Exception
      */
     public function testCanReadFailed()
     {
         $object = new Word2007();
-        $fqFilename = join(
-            DIRECTORY_SEPARATOR,
-            array(PHPWORD_TESTS_BASE_DIR, 'PhpWord', 'Tests', '_files', 'documents', 'foo.docx')
-        );
-        $this->assertFalse($object->canRead($fqFilename));
-        $object = IOFactory::load($fqFilename);
+        $filename = __DIR__ . '/../_files/documents/foo.docx';
+        $this->assertFalse($object->canRead($filename));
     }
 
     /**
      * Load
      */
     public function testLoad()
     {
-        $fqFilename = join(
-            DIRECTORY_SEPARATOR,
-            array(PHPWORD_TESTS_BASE_DIR, 'PhpWord', 'Tests', '_files', 'documents', 'reader.docx')
-        );
-        $object = IOFactory::load($fqFilename);
-        $this->assertInstanceOf('PhpOffice\\PhpWord\\PhpWord', $object);
+        $filename = __DIR__ . '/../_files/documents/reader.docx';
+        $phpWord = IOFactory::load($filename);
+        $this->assertInstanceOf('PhpOffice\\PhpWord\\PhpWord', $phpWord);
     }
 }
diff --git a/tests/PhpWord/Tests/_files/documents/reader.rtf b/tests/PhpWord/Tests/_files/documents/reader.rtf
@@ -0,0 +1,21 @@
+{\rtf1
+\ansi\ansicpg1252
+\deff0
+{\fonttbl{\f0\fnil\fcharset0 Arial;}{\f1\fnil\fcharset0 Times New Roman;}}
+{\colortbl;\red255\green0\blue0;\red14\green0\blue0}
+{\*\generator PhpWord;}
+
+{\info{\title }{\subject }{\category }{\keywords }{\comment }{\author }{\operator }{\creatim \yr2014\mo05\dy27\hr23\min36\sec45}{\revtim \yr2014\mo05\dy27\hr23\min36\sec45}{\company }{\manager }}
+\deftab720\viewkind1\uc1\pard\nowidctlpar\lang1036\kerning1\fs20
+{Welcome to PhpWord}\par
+\pard\nowidctlpar{\cf0\f0 Hello World!}\par
+\par
+\par
+\pard\nowidctlpar{\cf0\f0\fs32\b\i I am styled by a <font style> definition.}\par
+\pard\nowidctlpar{\cf0\f0 I am styled by a paragraph style definition.}\par
+\pard\nowidctlpar\qc\sa100{\cf0\f0\fs32\b\i I am styled by both font and paragraph style.}\par
+\pard\nowidctlpar{\cf1\f1\fs40\b\i\ul\strike\super I am inline styled.}\par
+\par
+{\field {\*\fldinst {HYPERLINK "http://www.google.com"}}{\fldrslt {Google}}}\par
+\par
+}

Original file line number	Diff line number	Diff line change
`@@ -33,7 +33,7 @@ class ODTextTest extends \PHPUnit_Framework_TestCase`
`33`	`33`	`public function testLoad()`
`34`	`34`	`{`
`35`	`35`	`$filename = __DIR__ . '/../_files/documents/reader.odt';`
`36`		`- $object = IOFactory::load($filename, 'ODText');`
`37`		`- $this->assertInstanceOf('PhpOffice\\PhpWord\\PhpWord', $object);`
	`36`	`+ $phpWord = IOFactory::load($filename, 'ODText');`
	`37`	`+ $this->assertInstanceOf('PhpOffice\\PhpWord\\PhpWord', $phpWord);`
`38`	`38`	`}`
`39`	`39`	`}`