Merge pull request #35 from DataValues/trimNormalizer

Add TrimmingStringNormalizer

README.md

@@ -57,6 +57,7 @@ employees for the [Wikidata project](https://wikidata.org/).
 
 ### 1.0.0 (dev)
 
+* Added `TrimmingStringNormalizer`
 * Removed the `DATAVALUES_COMMON_VERSION` constant
 * The `StringFormatter` constructor does not accept options any more
 

src/ValueParsers/Normalizers/TrimmingStringNormalizer.php

@@ -0,0 +1,32 @@
+<?php
+
+namespace ValueParsers\Normalizers;
+
+use InvalidArgumentException;
+
+/**
+ * Most simple implementation of a StringNormalizer that does nothing but trimming whitespace, by
+ * using the definition of "whitespace" in Unicode-enabled Perl regular expressions.
+ *
+ * @since 0.3.2
+ *
+ * @license GPL-2.0-or-later
+ * @author Thiemo Kreuz
+ */
+class TrimmingStringNormalizer implements StringNormalizer {
+
+	/**
+	 * @param string $value
+	 *
+	 * @throws InvalidArgumentException if $value is not a string
+	 * @return string the trimmed string
+	 */
+	public function normalize( $value ) {
+		if ( !is_string( $value ) ) {
+			throw new InvalidArgumentException( '$value must be a string' );
+		}
+
+		return preg_replace( '/^\s+|\s+$/u', '', $value );
+	}
+
+}

tests/ValueParsers/Normalizers/TrimmingStringNormalizerTest.php

@@ -0,0 +1,56 @@
+<?php
+
+namespace ValueParsers\Normalizers\Test;
+
+use DataValues\StringValue;
+use PHPUnit_Framework_TestCase;
+use ValueParsers\Normalizers\TrimmingStringNormalizer;
+
+/**
+ * @covers ValueParsers\Normalizers\TrimmingStringNormalizer
+ *
+ * @group ValueParsers
+ * @group DataValueExtensions
+ *
+ * @license GPL-2.0-or-later
+ * @author Thiemo Kreuz
+ */
+class TrimmingStringNormalizerTest extends PHPUnit_Framework_TestCase {
+
+	/**
+	 * @dataProvider stringProvider
+	 */
+	public function testNormalize( $value, $expected ) {
+		$normalizer = new TrimmingStringNormalizer();
+		$this->assertSame( $expected, $normalizer->normalize( $value ) );
+	}
+
+	public function stringProvider() {
+		return [
+			'Empty' => [ '', '' ],
+			'Trimmed' => [ 'a', 'a' ],
+			'Spaces' => [ ' a ', 'a' ],
+			'Controls' => [ "\n\r\ta\n\r\t", 'a' ],
+			'Paragraph separator' => [ "\xE2\x80\xA9a\xE2\x80\xA9", 'a' ],
+		];
+	}
+
+	/**
+	 * @dataProvider invalidValueProvider
+	 */
+	public function testNormalizeException( $value ) {
+		$normalizer = new TrimmingStringNormalizer();
+		$this->setExpectedException( 'InvalidArgumentException' );
+		$normalizer->normalize( $value );
+	}
+
+	public function invalidValueProvider() {
+		return [
+			[ null ],
+			[ true ],
+			[ 1 ],
+			[ new StringValue( '' ) ],
+		];
+	}
+
+}