ElasticaUtil::getPhraseSuggestion() - Code Metrics - Inspection of "WIP: Trying to debug and fix tests on Travis" - gordonbanderson/silverstripe-elastica - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — dev2 ( e0c580...5942a1 )

by Gordon

created 2015-12-30 01:13 UTC

ElasticaUtil::getPhraseSuggestion() B

↳ Parent: ElasticaUtil

Complexity

Conditions	6
Paths	2

Size

Total Lines	86
Code Lines	51

Duplication

Lines	0
Ratio	0 %

Importance

Changes	1
Bugs	1	Features	0

Metric	Value
c	1
b	1
f	0
dl	0
loc	86
rs	8.3235
cc	6
eloc	51
nc	2
nop	1

How to fix Long Method

<?php

namespace SilverStripe\Elastica;

/**
 * Utility methods to help with searching functions, and also testable without fixtures
 */
class ElasticaUtil {

	/**
	 * Marker string for pre highlight - can be any string unlikely to appear in a search
	 */
	private static $pre_marker = " |PREZXCVBNM12345678";

	/**
	 * Marker string for psot highlight - can be any string unlikely to appear in a search
	 */
	private static $post_marker = "POSTZXCVBNM12345678| ";

	/**
	 * @var boolean true to show CLI output, false to hide
	 */
	private static $cli_printer_output = true;


	/**
	 * Function to display messages only if using the command line
	 * @var string $content Text to display when in command line mode
	 */
	public static function message($content) {
		if (\Director::is_cli() && self::$cli_printer_output == true) {

			echo "$content\n";
		}
	}


	/*
	Display a human readable yes or no
	 */
	public static function showBooleanHumanReadable($assertion) {
		return $assertion ? 'Yes' : 'No';
	}


	public static function getPhraseSuggestion($alternativeQuerySuggestions) {
		$originalQuery = $alternativeQuerySuggestions[0]['text'];

		$highlightsCfg = \Config::inst()->get('Elastica', 'Highlights');
		$preTags = $highlightsCfg['PreTags'];
		$postTags = $highlightsCfg['PostTags'];

		//Use the first suggested phrase
		$options = $alternativeQuerySuggestions[0]['options'];

		$resultArray = null;
		if (sizeof($options) > 0) {
			//take the first suggestion
			$suggestedPhrase = $options[0]['text'];
			$suggestedPhraseHighlighted = $options[0]['highlighted'];

			// now need to fix capitalisation
			$originalParts = explode(' ', $originalQuery);
			$suggestedParts = explode(' ', $suggestedPhrase);

			$markedHighlightedParts = ' '.$suggestedPhraseHighlighted.' ';
			$markedHighlightedParts = str_replace(' '.$preTags, ' '.self::$pre_marker, $markedHighlightedParts);

			$markedHighlightedParts = str_replace($postTags.' ', self::$post_marker, $markedHighlightedParts);

			$markedHighlightedParts = trim($markedHighlightedParts);
			$markedHighlightedParts = trim($markedHighlightedParts);

			$highlightedParts = preg_split('/\s+/', $markedHighlightedParts);

			//Create a mapping of lowercase to uppercase terms
			$lowerToUpper = array();
			$lowerToHighlighted = array();
			$ctr = 0;
			foreach ($suggestedParts as $lowercaseWord) {
				$lowerToUpper[$lowercaseWord] = $originalParts[$ctr];
				$lowerToHighlighted[$lowercaseWord] = $highlightedParts[$ctr];
				$ctr++;
			}

			$plain = array();
			$highlighted = array();
			foreach ($suggestedParts as $lowercaseWord) {
				$possiblyUppercase = $lowerToUpper[$lowercaseWord];
				$possiblyUppercaseHighlighted = $lowerToHighlighted[$lowercaseWord];

				//If the terms are identical other than case, e.g. new => New, then simply swap
				if (strtolower($possiblyUppercase) == $lowercaseWord) {
					array_push($plain, $possiblyUppercase);
					array_push($highlighted, $possiblyUppercase);
				} else {
					//Need to check capitalisation of terms suggested that are different

					$chr = mb_substr ($possiblyUppercase, 0, 1, "UTF-8");
    				if (mb_strtolower($chr, "UTF-8") != $chr) {
    					$upperLowercaseWord = $lowercaseWord;
    					$upperLowercaseWord[0] = $chr;

    					//$possiblyUppercaseHighlighted = str_replace($lowercaseWord, $possiblyUppercase, $possiblyUppercaseHighlighted);

    					$withHighlights = str_replace($lowercaseWord, $upperLowercaseWord, $possiblyUppercaseHighlighted);

    					$lowercaseWord[0] = $chr;

    					//str_replace(search, replace, subject)


    					array_push($plain, $lowercaseWord);
    					array_push($highlighted, $withHighlights);
    				} else {
    					//No need to capitalise, so add suggested word
    					array_push($plain, $lowercaseWord);

    					//No need to capitalise, so add suggested highlighted word
    					array_push($highlighted, $possiblyUppercaseHighlighted);
    				}
				}
			}

			$highlighted = ' '.implode(' ', $highlighted).' ';
			$highlighted = str_replace(self::$pre_marker, ' '.$preTags, $highlighted);
			$highlighted = str_replace(self::$post_marker, $postTags.' ', $highlighted);

			$resultArray['suggestedQuery'] = implode(' ', $plain);
			$resultArray['suggestedQueryHighlighted'] = trim($highlighted);
		}
		return $resultArray;
	}


	/**
	 * The output format of this function is not documented, so at best this is guess work to an
	 * extent.  Possible formats are:
	 * - ((Title.standard:great Content.standard:ammunition Content.standard:could
	 *     Content.standard:bair Content.standard:dancing Content.standard:column
	 *     Content.standard:company Content.standard:infantry Content.standard:men
	 *     Content.standard:soldier Content.standard:brigade Content.standard:zealand
	 *     Content.standard:new)~3)
	 *     -ConstantScore(_uid:GutenbergBookExtract#1519)
	 *  (Description: bay Description: mannerstram)
	 *
	 * @param  string $explanation explanation string for more like this terms from Elasticsearch
	 * @return array             Array of fieldnames mapped to terms
	 */
	public static function parseSuggestionExplanation($explanation) {

		$explanation = explode('-ConstantScore', $explanation)[0];

        $bracketPos = strpos($explanation, ')~');

        if (substr($explanation, 0,2) == '((') {
        	$explanation = substr($explanation, 2, $bracketPos-2);
        } elseif (substr($explanation, 0,1) == '(') {
        	$explanation = substr($explanation, 1, $bracketPos-2);
        }

       	$terms = array();

        //Field name(s) => terms
        $splits = explode(' ', $explanation);

        foreach ($splits as $fieldAndTerm) {
        	$splits = explode(':', $fieldAndTerm);

        	// This is the no terms case
        	if (sizeof($splits) < 2) {
        		break;
        	}

        	$fieldname = $splits[0];
        	$term = $splits[1];

        	if (!isset($terms[$fieldname])) {
        		$terms[$fieldname] = array();
        	}

        	array_push($terms[$fieldname], $term);
        }

        return $terms;
	}

	/**
	 * Add attributes necessary for jQuery to execute autocomplete
	 * @param FormField &$queryField field used to type a search query
	 */
	public static function addAutocompleteToQueryField(&$queryField, $classesToSearch, $siteTreeOnly, $link, $slug) {
			$queryField->setAttribute('data-autocomplete', 'true');
			$queryField->setAttribute('data-autocomplete-field', 'Title');
			$queryField->setAttribute('data-autocomplete-classes', $classesToSearch);
			$queryField->setAttribute('data-autocomplete-sitetree', $siteTreeOnly);
			$queryField->setAttribute('data-autocomplete-source', $link);
			$queryField->setAttribute('data-autocomplete-function', $slug);
	}

	/**
	 * @return function print content to either web browser or command line.  Can be optionally supressed
	 */
	public static function getPrinter() {
		return function ($content) {
			if (self::$cli_printer_output == true) {

				print(\Director::is_cli() ? "T1 $content\n" : "T2 <p>$content</p>");
			}

		};
	}

	/**
	 * Set to true to show output on the command line or browser, false to not
	 *
	 * @param   $newcli_printer_output true to show output, false to hide it
	 */
	public static function setPrinterOutput($new_cli_printer_output) {

		self::$cli_printer_output = $new_cli_printer_output;
	}
}


1			<?php
2
3			namespace SilverStripe\Elastica;
4
5			/**
6			* Utility methods to help with searching functions, and also testable without fixtures
7			*/
8			class ElasticaUtil {
9
10			/**
11			* Marker string for pre highlight - can be any string unlikely to appear in a search
12			*/
13			private static $pre_marker = " \|PREZXCVBNM12345678";
14
15			/**
16			* Marker string for psot highlight - can be any string unlikely to appear in a search
17			*/
18			private static $post_marker = "POSTZXCVBNM12345678\| ";
19
20			/**
21			* @var boolean true to show CLI output, false to hide
22			*/
23			private static $cli_printer_output = true;
24
25
26			/**
27			* Function to display messages only if using the command line
28			* @var string $content Text to display when in command line mode
29			*/
30			public static function message($content) {
31			if (\Director::is_cli() && self::$cli_printer_output == true) {
			0 ignored issues – show Coding Style Best Practice introduced 2015-12-30 01:25 UTC by Report Bug Copy Issue Report It seems like you are loosely comparing two booleans. Considering using the strict comparison `===` instead. When comparing two booleans, it is generally considered safer to use the strict comparison operator. Loading history...
32			echo "$content\n";
33			}
34			}
35
36
37			/*
38			Display a human readable yes or no
39			*/
40			public static function showBooleanHumanReadable($assertion) {
41			return $assertion ? 'Yes' : 'No';
42			}
43
44
45			public static function getPhraseSuggestion($alternativeQuerySuggestions) {
46			$originalQuery = $alternativeQuerySuggestions[0]['text'];
47
48			$highlightsCfg = \Config::inst()->get('Elastica', 'Highlights');
49			$preTags = $highlightsCfg['PreTags'];
50			$postTags = $highlightsCfg['PostTags'];
51
52			//Use the first suggested phrase
53			$options = $alternativeQuerySuggestions[0]['options'];
54
55			$resultArray = null;
56			if (sizeof($options) > 0) {
57			//take the first suggestion
58			$suggestedPhrase = $options[0]['text'];
59			$suggestedPhraseHighlighted = $options[0]['highlighted'];
60
61			// now need to fix capitalisation
62			$originalParts = explode(' ', $originalQuery);
63			$suggestedParts = explode(' ', $suggestedPhrase);
64
65			$markedHighlightedParts = ' '.$suggestedPhraseHighlighted.' ';
66			$markedHighlightedParts = str_replace(' '.$preTags, ' '.self::$pre_marker, $markedHighlightedParts);
67
68			$markedHighlightedParts = str_replace($postTags.' ', self::$post_marker, $markedHighlightedParts);
69
70			$markedHighlightedParts = trim($markedHighlightedParts);
71			$markedHighlightedParts = trim($markedHighlightedParts);
72
73			$highlightedParts = preg_split('/\s+/', $markedHighlightedParts);
74
75			//Create a mapping of lowercase to uppercase terms
76			$lowerToUpper = array();
77			$lowerToHighlighted = array();
78			$ctr = 0;
79			foreach ($suggestedParts as $lowercaseWord) {
80			$lowerToUpper[$lowercaseWord] = $originalParts[$ctr];
81			$lowerToHighlighted[$lowercaseWord] = $highlightedParts[$ctr];
82			$ctr++;
83			}
84
85			$plain = array();
86			$highlighted = array();
87			foreach ($suggestedParts as $lowercaseWord) {
88			$possiblyUppercase = $lowerToUpper[$lowercaseWord];
89			$possiblyUppercaseHighlighted = $lowerToHighlighted[$lowercaseWord];
90
91			//If the terms are identical other than case, e.g. new => New, then simply swap
92			if (strtolower($possiblyUppercase) == $lowercaseWord) {
93			array_push($plain, $possiblyUppercase);
94			array_push($highlighted, $possiblyUppercase);
95			} else {
96			//Need to check capitalisation of terms suggested that are different
97
98			$chr = mb_substr ($possiblyUppercase, 0, 1, "UTF-8");
99			if (mb_strtolower($chr, "UTF-8") != $chr) {
100			$upperLowercaseWord = $lowercaseWord;
101			$upperLowercaseWord[0] = $chr;
102
103			//$possiblyUppercaseHighlighted = str_replace($lowercaseWord, $possiblyUppercase, $possiblyUppercaseHighlighted);
			0 ignored issues – show Unused Code Comprehensibility introduced 2015-12-01 14:39 UTC by Report Bug Copy Issue Report `60%` of this comment could be valid code. Did you maybe forget this after debugging? Sometimes obsolete code just ends up commented out instead of removed. In this case it is better to remove the code once you have checked you do not need it. The code might also have been commented out for debugging purposes. In this case it is vital that someone uncomments it again or your project may behave in very unexpected ways in production. This check looks for comments that seem to be mostly valid code and reports them. Loading history...
104			$withHighlights = str_replace($lowercaseWord, $upperLowercaseWord, $possiblyUppercaseHighlighted);
105
106			$lowercaseWord[0] = $chr;
107
108			//str_replace(search, replace, subject)
			0 ignored issues – show Unused Code Comprehensibility introduced 2015-12-01 14:39 UTC by Report Bug Copy Issue Report `40%` of this comment could be valid code. Did you maybe forget this after debugging? Sometimes obsolete code just ends up commented out instead of removed. In this case it is better to remove the code once you have checked you do not need it. The code might also have been commented out for debugging purposes. In this case it is vital that someone uncomments it again or your project may behave in very unexpected ways in production. This check looks for comments that seem to be mostly valid code and reports them. Loading history...
109
110			array_push($plain, $lowercaseWord);
111			array_push($highlighted, $withHighlights);
112			} else {
113			//No need to capitalise, so add suggested word
114			array_push($plain, $lowercaseWord);
115
116			//No need to capitalise, so add suggested highlighted word
117			array_push($highlighted, $possiblyUppercaseHighlighted);
118			}
119			}
120			}
121
122			$highlighted = ' '.implode(' ', $highlighted).' ';
123			$highlighted = str_replace(self::$pre_marker, ' '.$preTags, $highlighted);
124			$highlighted = str_replace(self::$post_marker, $postTags.' ', $highlighted);
125
126			$resultArray['suggestedQuery'] = implode(' ', $plain);
127			$resultArray['suggestedQueryHighlighted'] = trim($highlighted);
128			}
129			return $resultArray;
130			}
131
132
133			/**
134			* The output format of this function is not documented, so at best this is guess work to an
135			* extent. Possible formats are:
136			* - ((Title.standard:great Content.standard:ammunition Content.standard:could
137			* Content.standard:bair Content.standard:dancing Content.standard:column
138			* Content.standard:company Content.standard:infantry Content.standard:men
139			* Content.standard:soldier Content.standard:brigade Content.standard:zealand
140			* Content.standard:new)~3)
141			* -ConstantScore(_uid:GutenbergBookExtract#1519)
142			* (Description: bay Description: mannerstram)
143			*
144			* @param string $explanation explanation string for more like this terms from Elasticsearch
145			* @return array Array of fieldnames mapped to terms
146			*/
147			public static function parseSuggestionExplanation($explanation) {
148
149			$explanation = explode('-ConstantScore', $explanation)[0];
150
151			$bracketPos = strpos($explanation, ')~');
152
153			if (substr($explanation, 0,2) == '((') {
154			$explanation = substr($explanation, 2, $bracketPos-2);
155			} elseif (substr($explanation, 0,1) == '(') {
156			$explanation = substr($explanation, 1, $bracketPos-2);
157			}
158
159			$terms = array();
160
161			//Field name(s) => terms
162			$splits = explode(' ', $explanation);
163
164			foreach ($splits as $fieldAndTerm) {
165			$splits = explode(':', $fieldAndTerm);
166
167			// This is the no terms case
168			if (sizeof($splits) < 2) {
169			break;
170			}
171
172			$fieldname = $splits[0];
173			$term = $splits[1];
174
175			if (!isset($terms[$fieldname])) {
176			$terms[$fieldname] = array();
177			}
178
179			array_push($terms[$fieldname], $term);
180			}
181
182			return $terms;
183			}
184
185			/**
186			* Add attributes necessary for jQuery to execute autocomplete
187			* @param FormField &$queryField field used to type a search query
188			*/
189			public static function addAutocompleteToQueryField(&$queryField, $classesToSearch, $siteTreeOnly, $link, $slug) {
190			$queryField->setAttribute('data-autocomplete', 'true');
191			$queryField->setAttribute('data-autocomplete-field', 'Title');
192			$queryField->setAttribute('data-autocomplete-classes', $classesToSearch);
193			$queryField->setAttribute('data-autocomplete-sitetree', $siteTreeOnly);
194			$queryField->setAttribute('data-autocomplete-source', $link);
195			$queryField->setAttribute('data-autocomplete-function', $slug);
196			}
197
198			/**
199			* @return function print content to either web browser or command line. Can be optionally supressed
200			*/
201			public static function getPrinter() {
202			return function ($content) {
203			if (self::$cli_printer_output == true) {
			0 ignored issues – show Coding Style Best Practice introduced 2015-12-30 01:25 UTC by Report Bug Copy Issue Report It seems like you are loosely comparing two booleans. Considering using the strict comparison `===` instead. When comparing two booleans, it is generally considered safer to use the strict comparison operator. Loading history...
204			print(\Director::is_cli() ? "T1 $content\n" : "T2 <p>$content</p>");
205			}
206
207			};
208			}
209
210			/**
211			* Set to true to show output on the command line or browser, false to not
212			*
213			* @param $newcli_printer_output true to show output, false to hide it
214			*/
215			public static function setPrinterOutput($new_cli_printer_output) {
			1 ignored issue – show Coding Style Naming introduced 2015-12-30 01:25 UTC by Report Bug Copy Issue Report The parameter $new_cli_printer_output is not named in camelCase. This check marks parameter names that have not been written in camelCase. In camelCase names are written without any punctuation, the start of each new word being marked by a capital letter. Thus the name database connection string becomes `databaseConnectionString`. Loading history...
216			self::$cli_printer_output = $new_cli_printer_output;
217			}
218			}
219

gordonbanderson / silverstripe-elastica

Push — dev2 ( e0c580...5942a1 )

ElasticaUtil::getPhraseSuggestion() B

Complexity

Size

Duplication

Importance

How to fix Long Method

Long Method

Duplication Side-by-Side

Filter issues like