StatisticsRepository::getTopKeyWordsWithOrWithoutHits() - Code Metrics - Inspection of "[TASK] Extract access to frontend user session (#1..." - timohund/ext-solr - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( e20b48...fb7724 )

by Timo

created 2017-08-16 15:24 UTC

getTopKeyWordsWithOrWithoutHits() A

↳ Parent: StatisticsRepository

Complexity

Conditions	2
Paths	2

Size

Total Lines	18
Code Lines	11

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	6
CRAP Score	2.0116

Importance

Changes

Metric	Value
dl	0
loc	18
ccs	6
cts	7
cp	0.8571
rs	9.4285
c	0
b	0
f	0
cc	2
eloc	11
nc	2
nop	4
crap	2.0116

<?php declare(strict_types = 1);
namespace ApacheSolrForTypo3\Solr\Domain\Search\Statistics;

/***************************************************************
 *  Copyright notice
 *
 *  (c) 2016 Thomas Hohn <[email protected]>
 *  All rights reserved
 *
 *  This script is part of the TYPO3 project. The TYPO3 project is
 *  free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2 of the License, or
 *  (at your option) any later version.
 *
 *  The GNU General Public License can be found at
 *  http://www.gnu.org/copyleft/gpl.html.
 *
 *  This script is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  This copyright notice MUST APPEAR in all copies of the script!
 ***************************************************************/

use ApacheSolrForTypo3\Solr\System\Records\AbstractRepository;
use TYPO3\CMS\Core\Database\Query\QueryBuilder;

/**
 * Calculates the SearchQueryStatistics
 *
 * @author Thomas Hohn <[email protected]>
 */
class StatisticsRepository extends AbstractRepository
{
    /**
     * @var string
     */
    protected $table = 'tx_solr_statistics';

    /**
     * Fetches must popular search keys words from the table tx_solr_statistics
     *
     * @param int $rootPageId
     * @param int $days number of days of history to query
     * @param int $limit
     * @return mixed
     */
    public function getSearchStatistics(int $rootPageId, int $days = 30, $limit = 10)
    {
        $now = time();
        $timeStart = (int)($now - 86400 * $days); // 86400 seconds/day
        $limit = (int)$limit;

        $statisticsRows = $this->getPreparedQueryBuilderForSearchStatisticsAndTopKeywords($rootPageId, $timeStart, $limit)
            ->execute()->fetchAll();

        $statisticsRows = $this->mergeRowsWithSameKeyword($statisticsRows);

        $sumCount = $statisticsRows['sumCount'];
        foreach ($statisticsRows as $statisticsRow) {
            $sumCount += $statisticsRow['count'];
        }

        $statisticsRows = array_map(function($row) use ($sumCount) {
            $row['percent'] = $row['count'] * 100 / $sumCount;
            return $row;
        }, $statisticsRows);

        return $statisticsRows;
    }

    /**
     * Returns prepared QueryBuilder for two purposes:
     * for getSearchStatistics() and getTopKeyWordsWithOrWithoutHits() methods
     *
     * @param int $rootPageId
     * @param int $timeStart
     * @param int $limit
     * @return QueryBuilder
     */
    protected function getPreparedQueryBuilderForSearchStatisticsAndTopKeywords(int $rootPageId, int $timeStart, int $limit) : QueryBuilder
    {
        $queryBuilder = $this->getQueryBuilder();
        $statisticsQueryBuilder = $queryBuilder
            ->select('keywords', 'num_found AS hits')
            ->add('select', $queryBuilder->expr()->count('keywords', 'count'), true)
            ->from($this->table)
            ->andWhere(
                $queryBuilder->expr()->gt('tstamp', $timeStart),
                $queryBuilder->expr()->eq('root_pid', $rootPageId)
            )
            ->groupBy('keywords', 'num_found')
            ->orderBy('count', 'DESC')
            ->addOrderBy('hits', 'DESC')
            ->addOrderBy('keywords', 'ASC')
            ->setMaxResults($limit);

        return $statisticsQueryBuilder;
    }

    /**
     * Find Top search keywords with results
     *
     * @param int $rootPageId
     * @param int $days number of days of history to query
     * @param int $limit
     * @return array
     */
    public function getTopKeyWordsWithHits(int $rootPageId, int $days = 30, int $limit = 10) : array
    {
        return $this->getTopKeyWordsWithOrWithoutHits($rootPageId, $days, $limit, false);
    }

    /**
     * Find Top search keywords without results
     *
     * @param int $rootPageId
     * @param int $days number of days of history to query
     * @param int $limit
     * @return array
     */
    public function getTopKeyWordsWithoutHits(int $rootPageId, int $days = 30, int $limit = 10) : array
    {
        return $this->getTopKeyWordsWithOrWithoutHits($rootPageId, $days, $limit, true);
    }

    /**
     * Find Top search keywords with or without results
     *
     * @param int $rootPageId
     * @param int $days number of days of history to query
     * @param int $limit
     * @param bool $withoutHits
     * @return array
     */
    protected function getTopKeyWordsWithOrWithoutHits(int $rootPageId, int $days = 30, int $limit = 10, bool $withoutHits = false) : array
    {
        $now = time();
        $timeStart = $now - 86400 * $days; // 86400 seconds/day

        $queryBuilder = $this->getPreparedQueryBuilderForSearchStatisticsAndTopKeywords($rootPageId, $timeStart, $limit);
        // Check if we want without or with hits
        if ($withoutHits === true) {
            $queryBuilder->andWhere($queryBuilder->expr()->eq('num_found', 0));
        } else {
            $queryBuilder->andWhere($queryBuilder->expr()->gt('num_found', 0));
        }

        $statisticsRows = $queryBuilder->execute()->fetchAll();
        $statisticsRows = $this->mergeRowsWithSameKeyword($statisticsRows);

        return $statisticsRows;
    }

    /**
     * This method groups rows with the same term and different count and hits
     * and calculates the average.
     *
     * @param array $statisticsRows
     * @return array
     */
    protected function mergeRowsWithSameKeyword(array $statisticsRows) : array
    {
        $result = [];
        foreach ($statisticsRows as $statisticsRow) {
            $term = html_entity_decode($statisticsRow['keywords'], ENT_QUOTES);

            $mergedRow = isset($result[$term]) ? $result[$term] : ['mergedrows' => 0, 'count' => 0];
            $mergedRow['mergedrows']++;

            // for the hits we need to take the average
            $avgHits = $this->getAverageFromField($mergedRow, $statisticsRow, 'hits');
            $mergedRow['hits'] = (int)$avgHits;

            // for the count we need to take the sum, because it's the sum of searches
            $mergedRow['count'] = $mergedRow['count'] + $statisticsRow['count'];

            $mergedRow['keywords'] = $term;
            $result[$term] = $mergedRow;
        }

        return array_values($result);
    }

    /**
     * Get number of queries over time
     *
     * @param int $rootPageId
     * @param int $days number of days of history to query
     * @param int $bucketSeconds Seconds per bucket
     * @return array [labels, data]
     */
    public function getQueriesOverTime(int $rootPageId, int $days = 30, int $bucketSeconds = 3600) : array
    {
        $now = time();
        $timeStart = $now - 86400 * intval($days); // 86400 seconds/day

        $queryBuilder = $this->getQueryBuilder();
        $result = $queryBuilder
            ->addSelectLiteral(
                'FLOOR(`tstamp`/' . $bucketSeconds . ') AS `bucket`',
                // @todo: Works only with MySQL. Add own column with Date type to prevent converting DateTime to Date
                'unix_timestamp(from_unixtime(`tstamp`, "%y-%m-%d")) AS `timestamp`',
                $queryBuilder->expr()->count('*', 'numQueries')
            )
            ->from($this->table)
            ->andWhere(
                $queryBuilder->expr()->gt('tstamp', $timeStart),
                $queryBuilder->expr()->eq('root_pid', $rootPageId)
            )
            ->groupBy('bucket', 'timestamp')
            ->orderBy('bucket', 'ASC')
            ->execute()->fetchAll();

        return $result;
    }

    /**
     * Regurns a result set by given plugin.tx_solr.search.frequentSearches.select configuration.
     *
     * @param array $frequentSearchConfiguration
     * @return array Array of frequent search terms, keys are the terms, values are hits
     */
    public function getFrequentSearchTermsFromStatisticsByFrequentSearchConfiguration(array $frequentSearchConfiguration) : array
    {
        $queryBuilder = $this->getQueryBuilder();
        $resultSet = $queryBuilder
            ->addSelectLiteral(
                $frequentSearchConfiguration['select.']['SELECT']
            )
            ->from($frequentSearchConfiguration['select.']['FROM'])
            ->add('where', $frequentSearchConfiguration['select.']['ADD_WHERE'], true)
            ->add('groupBy', $frequentSearchConfiguration['select.']['GROUP_BY'], true)
            ->add('orderBy', $frequentSearchConfiguration['select.']['ORDER_BY'])
            ->setMaxResults((int)$frequentSearchConfiguration['limit'])
            ->execute()->fetchAll();

        return $resultSet;
    }

    /**
     * This method is used to get an average value from merged statistic rows.
     *
     * @param array $mergedRow
     * @param array $statisticsRow
     * @param string $fieldName
     * @return float|int
     */
    protected function getAverageFromField(array &$mergedRow, array $statisticsRow, string $fieldName)
    {
        // when this is the first row we can take it.
        if ($mergedRow['mergedrows'] === 1) {
            $avgCount = $statisticsRow[$fieldName];
            return $avgCount;
        }

        $oldAverage = $mergedRow[$fieldName];
        $oldMergeRows = $mergedRow['mergedrows'] - 1;
        $oldCount = $oldAverage * $oldMergeRows;
        $avgCount = (($oldCount + $statisticsRow[$fieldName]) / $mergedRow['mergedrows']);
        return $avgCount;
    }
}


1		<?php declare(strict_types = 1);
2		namespace ApacheSolrForTypo3\Solr\Domain\Search\Statistics;
3
4		/***************************************************************
5		* Copyright notice
6		*
7		* (c) 2016 Thomas Hohn <[email protected]>
8		* All rights reserved
9		*
10		* This script is part of the TYPO3 project. The TYPO3 project is
11		* free software; you can redistribute it and/or modify
12		* it under the terms of the GNU General Public License as published by
13		* the Free Software Foundation; either version 2 of the License, or
14		* (at your option) any later version.
15		*
16		* The GNU General Public License can be found at
17		* http://www.gnu.org/copyleft/gpl.html.
18		*
19		* This script is distributed in the hope that it will be useful,
20		* but WITHOUT ANY WARRANTY; without even the implied warranty of
21		* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
22		* GNU General Public License for more details.
23		*
24		* This copyright notice MUST APPEAR in all copies of the script!
25		***************************************************************/
26
27		use ApacheSolrForTypo3\Solr\System\Records\AbstractRepository;
28		use TYPO3\CMS\Core\Database\Query\QueryBuilder;
29
30		/**
31		* Calculates the SearchQueryStatistics
32		*
33		* @author Thomas Hohn <[email protected]>
34		*/
35		class StatisticsRepository extends AbstractRepository
36		{
37		/**
38		* @var string
39		*/
40		protected $table = 'tx_solr_statistics';
41
42		/**
43	1	* Fetches must popular search keys words from the table tx_solr_statistics
44		*
45	1	* @param int $rootPageId
46	1	* @param int $days number of days of history to query
47	1	* @param int $limit
48	1	* @return mixed
49		*/
50	1	public function getSearchStatistics(int $rootPageId, int $days = 30, $limit = 10)
51	1	{
52	1	$now = time();
53	1	$timeStart = (int)($now - 86400 * $days); // 86400 seconds/day
54	1	$limit = (int)$limit;
55	1
56		$statisticsRows = $this->getPreparedQueryBuilderForSearchStatisticsAndTopKeywords($rootPageId, $timeStart, $limit)
57		->execute()->fetchAll();
58
59	1	$statisticsRows = $this->mergeRowsWithSameKeyword($statisticsRows);
60
61	1	$sumCount = $statisticsRows['sumCount'];
62	1	foreach ($statisticsRows as $statisticsRow) {
63		$sumCount += $statisticsRow['count'];
64		}
65
66	1	$statisticsRows = array_map(function($row) use ($sumCount) {
67		$row['percent'] = $row['count'] * 100 / $sumCount;
68		return $row;
69	1	}, $statisticsRows);
70
71	1	return $statisticsRows;
72		}
73
74		/**
75		* Returns prepared QueryBuilder for two purposes:
76		* for getSearchStatistics() and getTopKeyWordsWithOrWithoutHits() methods
77		*
78		* @param int $rootPageId
79		* @param int $timeStart
80		* @param int $limit
81		* @return QueryBuilder
82	1	*/
83		protected function getPreparedQueryBuilderForSearchStatisticsAndTopKeywords(int $rootPageId, int $timeStart, int $limit) : QueryBuilder
84	1	{
85		$queryBuilder = $this->getQueryBuilder();
86		$statisticsQueryBuilder = $queryBuilder
87		->select('keywords', 'num_found AS hits')
88		->add('select', $queryBuilder->expr()->count('keywords', 'count'), true)
89		->from($this->table)
90		->andWhere(
91		$queryBuilder->expr()->gt('tstamp', $timeStart),
92		$queryBuilder->expr()->eq('root_pid', $rootPageId)
93		)
94		->groupBy('keywords', 'num_found')
95	2	->orderBy('count', 'DESC')
96		->addOrderBy('hits', 'DESC')
97	2	->addOrderBy('keywords', 'ASC')
98		->setMaxResults($limit);
99
100		return $statisticsQueryBuilder;
101		}
102
103		/**
104		* Find Top search keywords with results
105		*
106		* @param int $rootPageId
107		* @param int $days number of days of history to query
108		* @param int $limit
109	3	* @return array
110		*/
111	3	public function getTopKeyWordsWithHits(int $rootPageId, int $days = 30, int $limit = 10) : array
112	3	{
113	3	return $this->getTopKeyWordsWithOrWithoutHits($rootPageId, $days, $limit, false);
114		}
115	3
116	3	/**
117		* Find Top search keywords without results
118		*
119	3	* @param int $rootPageId
120	2	* @param int $days number of days of history to query
121		* @param int $limit
122	1	* @return array
123		*/
124		public function getTopKeyWordsWithoutHits(int $rootPageId, int $days = 30, int $limit = 10) : array
125	3	{
126	3	return $this->getTopKeyWordsWithOrWithoutHits($rootPageId, $days, $limit, true);
127	3	}
128	3
129	3	/**
130	3	* Find Top search keywords with or without results
131		*
132		* @param int $rootPageId
133		* @param int $days number of days of history to query
134	3	* @param int $limit
135		* @param bool $withoutHits
136	3	* @return array
137		*/
138		protected function getTopKeyWordsWithOrWithoutHits(int $rootPageId, int $days = 30, int $limit = 10, bool $withoutHits = false) : array
139		{
140		$now = time();
141		$timeStart = $now - 86400 * $days; // 86400 seconds/day
142
143		$queryBuilder = $this->getPreparedQueryBuilderForSearchStatisticsAndTopKeywords($rootPageId, $timeStart, $limit);
144		// Check if we want without or with hits
145		if ($withoutHits === true) {
146	4	$queryBuilder->andWhere($queryBuilder->expr()->eq('num_found', 0));
147		} else {
148	4	$queryBuilder->andWhere($queryBuilder->expr()->gt('num_found', 0));
149	4	}
150	2
151		$statisticsRows = $queryBuilder->execute()->fetchAll();
152	2	$statisticsRows = $this->mergeRowsWithSameKeyword($statisticsRows);
153	2
154		return $statisticsRows;
155		}
156	2
157	2	/**
158		* This method groups rows with the same term and different count and hits
159		* and calculates the average.
160	2	*
161		* @param array $statisticsRows
162	2	* @return array
163	2	*/
164		protected function mergeRowsWithSameKeyword(array $statisticsRows) : array
165		{
166	4	$result = [];
167		foreach ($statisticsRows as $statisticsRow) {
168		$term = html_entity_decode($statisticsRow['keywords'], ENT_QUOTES);
169
170		$mergedRow = isset($result[$term]) ? $result[$term] : ['mergedrows' => 0, 'count' => 0];
171		$mergedRow['mergedrows']++;
172
173		// for the hits we need to take the average
174		$avgHits = $this->getAverageFromField($mergedRow, $statisticsRow, 'hits');
175		$mergedRow['hits'] = (int)$avgHits;
176
177		// for the count we need to take the sum, because it's the sum of searches
178		$mergedRow['count'] = $mergedRow['count'] + $statisticsRow['count'];
179
180		$mergedRow['keywords'] = $term;
181		$result[$term] = $mergedRow;
182		}
183
184		return array_values($result);
185		}
186
187		/**
188		* Get number of queries over time
189		*
190		* @param int $rootPageId
191		* @param int $days number of days of history to query
192		* @param int $bucketSeconds Seconds per bucket
193		* @return array [labels, data]
194		*/
195		public function getQueriesOverTime(int $rootPageId, int $days = 30, int $bucketSeconds = 3600) : array
196		{
197		$now = time();
198		$timeStart = $now - 86400 * intval($days); // 86400 seconds/day
199
200		$queryBuilder = $this->getQueryBuilder();
201	2	$result = $queryBuilder
202		->addSelectLiteral(
203		'FLOOR(`tstamp`/' . $bucketSeconds . ') AS `bucket`',
204	2	// @todo: Works only with MySQL. Add own column with Date type to prevent converting DateTime to Date
205	2	'unix_timestamp(from_unixtime(`tstamp`, "%y-%m-%d")) AS `timestamp`',
206	2	$queryBuilder->expr()->count('*', 'numQueries')
207		)
208		->from($this->table)
209	1	->andWhere(
210	1	$queryBuilder->expr()->gt('tstamp', $timeStart),
211	1	$queryBuilder->expr()->eq('root_pid', $rootPageId)
212	1	)
213	1	->groupBy('bucket', 'timestamp')
214		->orderBy('bucket', 'ASC')
215		->execute()->fetchAll();
216
217		return $result;
218		}
219	4
220		/**
221	4	* Regurns a result set by given plugin.tx_solr.search.frequentSearches.select configuration.
222		*
223		* @param array $frequentSearchConfiguration
224		* @return array Array of frequent search terms, keys are the terms, values are hits
225		*/
226		public function getFrequentSearchTermsFromStatisticsByFrequentSearchConfiguration(array $frequentSearchConfiguration) : array
227		{
228		$queryBuilder = $this->getQueryBuilder();
229		$resultSet = $queryBuilder
230		->addSelectLiteral(
231		$frequentSearchConfiguration['select.']['SELECT']
232		)
233		->from($frequentSearchConfiguration['select.']['FROM'])
234		->add('where', $frequentSearchConfiguration['select.']['ADD_WHERE'], true)
235		->add('groupBy', $frequentSearchConfiguration['select.']['GROUP_BY'], true)
236		->add('orderBy', $frequentSearchConfiguration['select.']['ORDER_BY'])
237		->setMaxResults((int)$frequentSearchConfiguration['limit'])
238		->execute()->fetchAll();
239
240		return $resultSet;
241		}
242
243		/**
244		* This method is used to get an average value from merged statistic rows.
245		*
246		* @param array $mergedRow
247		* @param array $statisticsRow
248		* @param string $fieldName
249		* @return float\|int
250		*/
251		protected function getAverageFromField(array &$mergedRow, array $statisticsRow, string $fieldName)
252		{
253		// when this is the first row we can take it.
254		if ($mergedRow['mergedrows'] === 1) {
255		$avgCount = $statisticsRow[$fieldName];
256		return $avgCount;
257		}
258
259		$oldAverage = $mergedRow[$fieldName];
260		$oldMergeRows = $mergedRow['mergedrows'] - 1;
261		$oldCount = $oldAverage * $oldMergeRows;
262		$avgCount = (($oldCount + $statisticsRow[$fieldName]) / $mergedRow['mergedrows']);
263		return $avgCount;
264		}
265		}
266

timohund / ext-solr

Push — master ( e20b48...fb7724 )

getTopKeyWordsWithOrWithoutHits() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like