Twitter tokenizer

src/Tokenizers/TwitterTokenizer.php

@@ -0,0 +1,144 @@
+<?php
+
+
+namespace TextAnalysis\Tokenizers;
+
+/**
+ * Tokenize tweets, based on http://www.nltk.org/_modules/nltk/tokenize/casual.html
+ * @author yooper
+ */
+class TwitterTokenizer extends TokenizerAbstract
+{       
+    public function tokenize($string): array 
+    {
+        $matches = [];
+        $found = preg_match_all("~{$this->getJoinedRegexes()}~uim", $string, $matches);
+        if($found === 0) {
+            return [];
+        }
+        return $matches[0];
+    }
+
+    /**
+     * Return an array of regexes, ordered by priority
+     */
+    public function getRegexes() : array
+    {
+        return [
+            $this->getPhoneNumbers(),                        
+            $this->getUrls(),            
+            $this->getEmoticons(),            
+            $this->getHashTags(),           
+            $this->getAsciiArrows(),            
+            $this->getUsernames(),
+            $this->getHashTags(),            
+            $this->getWordsWith(),
+            $this->getNumbers(),            
+            $this->getWordsWithout(),            
+            $this->getEllipsisDots(),            
+            $this->getEverythingElse()
+        ];
+    }
+
+    /**
+     * 
+     * @return string
+     */
+    public function getJoinedRegexes() : string
+    {
+        return implode("|", $this->getRegexes());
+    }
+
+    public function getHtmlTags() : string
+    {
+        return '<[^>\s]+>';
+    }
+
+    public function getAsciiArrows() : string
+    {
+        return '[\-]+>|<[\-]+';
+    }
+
+    public function getUsernames() : string
+    {
+        return '(?:@[\w_]+)';
+    }
+
+    public function getHashTags() : string
+    {
+        return '(?:\#+[\w_]+[\w\'_\-]*[\w_]+)';
+    }
+
+    /**
+     * # Words with apostrophes or dashes.
+     * @return string
+     */
+    public function getWordsWith() : string
+    {
+        return "(?:[^\W\d_](?:[^\W\d_]|['\-_])+[^\W\d_])";
+    }
+
+    public function getEmailAddress() : string
+    {
+        return '[\w.+-]+@[\w-]+\.(?:[\w-]\.?)+[\w-]';
+    }
+
+    public function getNumbers() : string
+    {
+        return '(?:[+\-]?\d+[,/.:-]\d+[+\-]?)';
+    }
+
+    /**
+     * Words without apostrophes or dashes.
+     * @return string
+     */
+    public function getWordsWithout() : string
+    {
+        return '(?:[\w_]+)'; 
+    }
+
+    public function getEllipsisDots() : string
+    {
+        return '(?:\.(?:\s*\.){1,})';
+    }
+
+    /**
+     * Everything else that isn't whitespace.
+     * @return string
+     */
+    public function getEverythingElse() : string
+    {
+        return '(?:\S)';
+    }
+
+    /**
+     * Taken from https://stackoverflow.com/questions/2113908/what-regular-expression-will-match-valid-international-phone-numbers
+     * @return string
+     */
+    public function getPhoneNumbers() : string
+    {
+        return '(?:(?:\+?1\s*(?:[.-]\s*)?)?(?:\(\s*([2-9]1[02-9]|[2-9][02-8]1|[2-9][02-8][02-9])\s*\)|([2-9]1[02-9]|[2-9][02-8]1|[2-9][02-8][02-9]))\s*(?:[.-]\s*)?)?([2-9]1[02-9]|[2-9][02-9]1|[2-9][02-9]{2})\s*(?:[.-]\s*)?([0-9]{4})(?:\s*(?:#|x\.?|ext\.?|extension)\s*(\d+))?';
+    }
+
+    /**
+     * Taken from https://stackoverflow.com/questions/6427530/regular-expression-pattern-to-match-url-with-or-without-http-www
+     * @return string
+     */
+    public function getUrls() : string
+    {
+        $regex = '((https?|ftp)://)?'; // SCHEME
+        $regex .= '([a-z0-9+!*(),;?&=$_.-]+(:[a-z0-9+!*(),;?&=$_.-]+)?@)?'; // User and Pass
+        $regex .= '([a-z0-9\-\.]*)\.(([a-z]{2,4})|([0-9]{1,3}\.([0-9]{1,3})\.([0-9]{1,3})))'; // Host or IP
+        $regex .= "(:[0-9]{2,5})?"; // Port
+        $regex .= '(/([a-z0-9+$_%-]\.?)+)*/?'; // Path
+        $regex .= '(\?[a-z+&\$_.-][a-z0-9;:@&%=+/$_.-]*)?'; // GET Query
+        $regex .= '(#[a-z_.-][a-z0-9+$%_.-]*)?'; // Anchor
+        return $regex;
+    }
+
+    public function getEmoticons() : string
+    {           
+        return '(\:\w+\:|\<[\/\\]?3|[\(\)\\\D|\*\$][\-\^]?[\:\;\=]|[\:\;\=B8][\-\^]?[3DOPp\@\$\*\\\)\(\/\|])(?=\s|[\!\.\?]|$)';
+    }
+
+}

tests/TextAnalysis/Tokenizers/TwitterTokenizerTest.php

@@ -0,0 +1,25 @@
+<?php
+
+namespace Tests\TextAnalysis\Tokenizers;
+
+use TextAnalysis\Tokenizers\TwitterTokenizer;
+
+/**
+ *
+ * @author yooper
+ */
+class TwitterTokenizerTest extends \PHPUnit_Framework_TestCase
+{
+    public function testTokenizer()
+    {
+        $tokens = (new TwitterTokenizer)->tokenize('This is a common Tweet #format where @mentions and.errors!!!!like this:-))))) might #appear❤ ❤☺❤#ThisIsAHashtag!?!');
+        $this->assertCount(33, $tokens);
+
+    }
+
+    public function testForUrlAndEmail()
+    {
+        $tokens = (new TwitterTokenizer)->tokenize('Custom Software Development http://redbeardtechnologies.com/ 906-555-5555 or contact support at support@redbeardtechnologies.com :-)');
+        $this->assertCount(11, $tokens);        
+    }    
+}