PagesRepository::getWikidataInfo() - Code Metrics - Inspection of "ArticleInfo: moving wikidata and checkwiki queries..." - x-tools/xtools - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( f9a368...59d319 )

unknown

created 2017-07-09 00:15 UTC

PagesRepository::getWikidataInfo() B

↳ Parent: PagesRepository

Complexity

Conditions	2
Paths	2

Size

Total Lines	39
Code Lines	11

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
dl	0
loc	39
rs	8.8571
c	0
b	0
f	0
cc	2
eloc	11
nc	2
nop	1

<?php
/**
 * This file contains only the PagesRepository class.
 */

namespace Xtools;

use DateInterval;
use Mediawiki\Api\SimpleRequest;

/**
 * A PagesRepository fetches data about Pages, either singularly or for multiple.
 */
class PagesRepository extends Repository
{

    /**
     * Get metadata about a single page from the API.
     * @param Project $project The project to which the page belongs.
     * @param string $pageTitle Page title.
     * @return string[] Array with some of the following keys: pageid, title, missing, displaytitle,
     * url.
     */
    public function getPageInfo(Project $project, $pageTitle)
    {
        $info = $this->getPagesInfo($project, [$pageTitle]);
        return array_shift($info);
    }

    /**
     * Get metadata about a set of pages from the API.
     * @param Project $project The project to which the pages belong.
     * @param string[] $pageTitles Array of page titles.
     * @return string[] Array keyed by the page names, each element with some of the
     * following keys: pageid, title, missing, displaytitle, url.
     */
    public function getPagesInfo(Project $project, $pageTitles)
    {
        // @TODO: Also include 'extlinks' prop when we start checking for dead external links.
        $params = [
            'prop' => 'info|pageprops',
            'inprop' => 'protection|talkid|watched|watchers|notificationtimestamp|subjectid|url|readable|displaytitle',
            'converttitles' => '',
            // 'ellimit' => 20,

            // 'elexpandurl' => '',

            'titles' => join('|', $pageTitles),
            'formatversion' => 2
            // 'pageids' => $pageIds // FIXME: allow page IDs

        ];

        $query = new SimpleRequest('query', $params);
        $api = $this->getMediawikiApi($project);
        $res = $api->getRequest($query);
        $result = [];
        if (isset($res['query']['pages'])) {
            foreach ($res['query']['pages'] as $pageInfo) {
                $result[$pageInfo['title']] = $pageInfo;
            }
        }
        return $result;
    }

    /**
     * Get revisions of a single page.
     * @param Page $page The page.
     * @param User|null $user Specify to get only revisions by the given user.
     * @return string[] Each member with keys: id, timestamp, length-
     */
    public function getRevisions(Page $page, User $user = null)
    {
        $cacheKey = 'revisions.'.$page->getId();
        if ($user) {
            $cacheKey .= '.'.$user->getUsername();
        }

        if ($this->cache->hasItem($cacheKey)) {
            return $this->cache->getItem($cacheKey)->get();
        }

        $this->stopwatch->start($cacheKey, 'XTools');

        $revTable = $this->getTableName($page->getProject()->getDatabaseName(), 'revision');
        $userClause = $user ? "revs.rev_user_text in (:username) AND " : "";

        $query = "SELECT
                    revs.rev_id AS id,
                    revs.rev_timestamp AS timestamp,
                    revs.rev_minor_edit AS minor,
                    revs.rev_len AS length,
                    (CAST(revs.rev_len AS SIGNED) - IFNULL(parentrevs.rev_len, 0)) AS length_change,
                    revs.rev_user AS user_id,
                    revs.rev_user_text AS username,
                    revs.rev_comment AS comment
                FROM $revTable AS revs
                LEFT JOIN $revTable AS parentrevs ON (revs.rev_parent_id = parentrevs.rev_id)
                WHERE $userClause revs.rev_page = :pageid
                ORDER BY revs.rev_timestamp ASC
            ";
        $params = ['pageid' => $page->getId()];
        if ($user) {
            $params['username'] = $user->getUsername();
        }
        $conn = $this->getProjectsConnection();
        $result = $conn->executeQuery($query, $params)->fetchAll();

        // Cache for 10 minutes, and return.

        $cacheItem = $this->cache->getItem($cacheKey)
            ->set($result)
            ->expiresAfter(new DateInterval('PT10M'));
        $this->cache->save($cacheItem);
        $this->stopwatch->stop($cacheKey);
    }

    /**
     * Get a count of the number of revisions of a single page
     * @param Page $page The page.
     * @param User|null $user Specify to only count revisions by the given user.
     * @return int
     */
    public function getNumRevisions(Page $page, User $user = null)
    {
        $revTable = $this->getTableName($page->getProject()->getDatabaseName(), 'revision');
        $userClause = $user ? "rev_user_text in (:username) AND " : "";

        $query = "SELECT COUNT(*)
                FROM $revTable
                WHERE $userClause rev_page = :pageid
            ";
        $params = ['pageid' => $page->getId()];
        if ($user) {
            $params['username'] = $user->getUsername();
        }
        $conn = $this->getProjectsConnection();
        return $conn->executeQuery($query, $params)->fetchColumn(0);
    }

    /**
     * Get assessment data for the given pages
     * @param Project   $project The project to which the pages belong.
     * @param  int[]    $pageIds Page IDs
     * @return string[] Assessment data as retrieved from the database.
     */
    public function getAssessments(Project $project, $pageIds)
    {
        if (!$project->hasPageAssessments()) {
            return [];
        }
        $pageAssessmentsTable = $this->getTableName($project->getDatabaseName(), 'page_assessments');
$myVar = 'Value';
$higher = false;

if (rand(1, 6) > 3) {
    $higher = true;
} else {
    $higher = false;
}
        $pageIds = implode($pageIds, ',');

        $query = "SELECT pap_project_title AS wikiproject, pa_class AS class, pa_importance AS importance
                  FROM page_assessments
                  LEFT JOIN page_assessments_projects ON pa_project_id = pap_project_id
                  WHERE pa_page_id IN ($pageIds)";

        $conn = $this->getProjectsConnection();
        return $conn->executeQuery($query)->fetchAll();
    }

    /**
     * Get any CheckWiki errors of a single page
     * @param Page $page
     * @return array Results from query
     */
    public function getCheckWikiErrors(Page $page)
    {
        // Only support mainspace on Labs installations
        if ($page->getNamespace() !== 0 || !$this->isLabs()) {
            return [];
        }

        $sql = "SELECT error, notice, found, name_trans AS name, prio, text_trans AS explanation
                FROM s51080__checkwiki_p.cw_error a
                JOIN s51080__checkwiki_p.cw_overview_errors b
                WHERE a.project = b.project
                AND a.project = :dbName
                AND a.title = :title
                AND a.error = b.id
                AND a.ok = 0";

        // remove _p if present
        $dbName = preg_replace('/_p$/', '', $page->getProject()->getDatabaseName());

        // Page title without underscores (str_replace just to be sure)
        $pageTitle = str_replace('_', ' ', $page->getTitle());

        $resultQuery = $this->getToolsConnection()->prepare($sql);
        $resultQuery->bindParam(':dbName', $dbName);
        $resultQuery->bindParam(':title', $pageTitle);
        $resultQuery->execute();

        return $resultQuery->fetchAll();
    }

    /**
     * Get basic wikidata on the page: label and description.
     * @param Page $page
     * @return string[] In the format:
     *    [[
     *         'term' => string such as 'label',
     *         'term_text' => string (value for 'label'),
     *     ], ... ]
     */
    public function getWikidataInfo(Page $page)
    {
        if (empty($page->getWikidataId())) {
            return [];
        }

        $wikidataId = ltrim($page->getWikidataId(), 'Q');
        $lang = $page->getProject()->getLang();

        $sql = "SELECT IF(term_type = 'label', 'label', 'description') AS term, term_text
                FROM wikidatawiki_p.wb_entity_per_page
                JOIN wikidatawiki_p.page ON epp_page_id = page_id
                JOIN wikidatawiki_p.wb_terms ON term_entity_id = epp_entity_id
                    AND term_language = :lang
                    AND term_type IN ('label', 'description')
                WHERE epp_entity_id = :wikidataId

                UNION

                SELECT pl_title AS term, wb_terms.term_text
                FROM wikidatawiki_p.pagelinks
                JOIN wikidatawiki_p.wb_terms ON term_entity_id = SUBSTRING(pl_title, 2)
                    AND term_entity_type = (IF(SUBSTRING(pl_title, 1, 1) = 'Q', 'item', 'property'))
                    AND term_language = :lang
                    AND term_type = 'label'
                WHERE pl_namespace IN (0, 120)
                    AND pl_from = (
                        SELECT page_id FROM page
                        WHERE page_namespace = 0
                            AND page_title = 'Q:wikidataId'
                    )";

        $resultQuery = $this->getProjectsConnection()->prepare($sql);
        $resultQuery->bindParam(':lang', $lang);
        $resultQuery->bindParam(':wikidataId', $wikidataId);
        $resultQuery->execute();

        return $resultQuery->fetchAll();
    }

    /**
     * Get or count all wikidata items for the given page,
     *     not just languages of sister projects
     * @param Page $page
     * @param bool $count Set to true to get only a COUNT
     * @return string[]|int Records as returend by the DB,
     *                      or raw COUNT of the records.
     */
    public function getWikidataItems(Page $page, $count = false)
    {
        if (!$page->getWikidataId()) {
            return $count ? 0 : [];
        }

        $wikidataId = ltrim($page->getWikidataId(), 'Q');

        $sql = "SELECT " . ($count ? 'COUNT(*) AS count' : '*') . "
                FROM wikidatawiki_p.wb_items_per_site
                WHERE ips_item_id = :wikidataId";

        $resultQuery = $this->getProjectsConnection()->prepare($sql);
        $resultQuery->bindParam(':wikidataId', $wikidataId);
        $resultQuery->execute();

        $result = $resultQuery->fetchAll();

        return $count ? (int) $result[0]['count'] : $result;
    }

    /**
     * Count wikidata items for the given page, not just languages of sister projects
     * @param Page $page
     * @return int Number of records.
     */
    public function countWikidataItems(Page $page)
    {
        return $this->getWikidataItems($page, true);
    }
}


1			<?php
2			/**
3			* This file contains only the PagesRepository class.
4			*/
5
6			namespace Xtools;
7
8			use DateInterval;
9			use Mediawiki\Api\SimpleRequest;
10
11			/**
12			* A PagesRepository fetches data about Pages, either singularly or for multiple.
13			*/
14			class PagesRepository extends Repository
15			{
16
17			/**
18			* Get metadata about a single page from the API.
19			* @param Project $project The project to which the page belongs.
20			* @param string $pageTitle Page title.
21			* @return string[] Array with some of the following keys: pageid, title, missing, displaytitle,
22			* url.
23			*/
24			public function getPageInfo(Project $project, $pageTitle)
25			{
26			$info = $this->getPagesInfo($project, [$pageTitle]);
27			return array_shift($info);
28			}
29
30			/**
31			* Get metadata about a set of pages from the API.
32			* @param Project $project The project to which the pages belong.
33			* @param string[] $pageTitles Array of page titles.
34			* @return string[] Array keyed by the page names, each element with some of the
35			* following keys: pageid, title, missing, displaytitle, url.
36			*/
37			public function getPagesInfo(Project $project, $pageTitles)
38			{
39			// @TODO: Also include 'extlinks' prop when we start checking for dead external links.
40			$params = [
41			'prop' => 'info\|pageprops',
42			'inprop' => 'protection\|talkid\|watched\|watchers\|notificationtimestamp\|subjectid\|url\|readable\|displaytitle',
43			'converttitles' => '',
44			// 'ellimit' => 20,
			1 ignored issue – show Unused Code Comprehensibility introduced 2017-03-23 00:29 UTC by Report Bug Copy Issue Report `58%` 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...
45			// 'elexpandurl' => '',
			1 ignored issue – show Unused Code Comprehensibility introduced 2017-03-23 00:29 UTC by Report Bug Copy Issue Report `58%` 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...
46			'titles' => join('\|', $pageTitles),
47			'formatversion' => 2
48			// 'pageids' => $pageIds // FIXME: allow page IDs
			1 ignored issue – show Unused Code Comprehensibility introduced 2017-03-23 00:29 UTC by Report Bug Copy Issue Report `43%` 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...
49			];
50
51			$query = new SimpleRequest('query', $params);
52			$api = $this->getMediawikiApi($project);
53			$res = $api->getRequest($query);
54			$result = [];
55			if (isset($res['query']['pages'])) {
56			foreach ($res['query']['pages'] as $pageInfo) {
57			$result[$pageInfo['title']] = $pageInfo;
58			}
59			}
60			return $result;
61			}
62
63			/**
64			* Get revisions of a single page.
65			* @param Page $page The page.
66			* @param User\|null $user Specify to get only revisions by the given user.
67			* @return string[] Each member with keys: id, timestamp, length-
68			*/
69			public function getRevisions(Page $page, User $user = null)
70			{
71			$cacheKey = 'revisions.'.$page->getId();
72			if ($user) {
73			$cacheKey .= '.'.$user->getUsername();
74			}
75
76			if ($this->cache->hasItem($cacheKey)) {
77			return $this->cache->getItem($cacheKey)->get();
78			}
79
80			$this->stopwatch->start($cacheKey, 'XTools');
81
82			$revTable = $this->getTableName($page->getProject()->getDatabaseName(), 'revision');
83			$userClause = $user ? "revs.rev_user_text in (:username) AND " : "";
84
85			$query = "SELECT
86			revs.rev_id AS id,
87			revs.rev_timestamp AS timestamp,
88			revs.rev_minor_edit AS minor,
89			revs.rev_len AS length,
90			(CAST(revs.rev_len AS SIGNED) - IFNULL(parentrevs.rev_len, 0)) AS length_change,
91			revs.rev_user AS user_id,
92			revs.rev_user_text AS username,
93			revs.rev_comment AS comment
94			FROM $revTable AS revs
95			LEFT JOIN $revTable AS parentrevs ON (revs.rev_parent_id = parentrevs.rev_id)
96			WHERE $userClause revs.rev_page = :pageid
97			ORDER BY revs.rev_timestamp ASC
98			";
99			$params = ['pageid' => $page->getId()];
100			if ($user) {
101			$params['username'] = $user->getUsername();
102			}
103			$conn = $this->getProjectsConnection();
104			$result = $conn->executeQuery($query, $params)->fetchAll();
105
106			// Cache for 10 minutes, and return.
			1 ignored issue – show Unused Code Comprehensibility introduced 2017-05-16 01:27 UTC by Report Bug Copy Issue Report `36%` 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...
107			$cacheItem = $this->cache->getItem($cacheKey)
108			->set($result)
109			->expiresAfter(new DateInterval('PT10M'));
110			$this->cache->save($cacheItem);
111			$this->stopwatch->stop($cacheKey);
112			}
113
114			/**
115			* Get a count of the number of revisions of a single page
116			* @param Page $page The page.
117			* @param User\|null $user Specify to only count revisions by the given user.
118			* @return int
119			*/
120			public function getNumRevisions(Page $page, User $user = null)
121			{
122			$revTable = $this->getTableName($page->getProject()->getDatabaseName(), 'revision');
123			$userClause = $user ? "rev_user_text in (:username) AND " : "";
124
125			$query = "SELECT COUNT(*)
126			FROM $revTable
127			WHERE $userClause rev_page = :pageid
128			";
129			$params = ['pageid' => $page->getId()];
130			if ($user) {
131			$params['username'] = $user->getUsername();
132			}
133			$conn = $this->getProjectsConnection();
134			return $conn->executeQuery($query, $params)->fetchColumn(0);
135			}
136
137			/**
138			* Get assessment data for the given pages
139			* @param Project $project The project to which the pages belong.
140			* @param int[] $pageIds Page IDs
141			* @return string[] Assessment data as retrieved from the database.
142			*/
143			public function getAssessments(Project $project, $pageIds)
144			{
145			if (!$project->hasPageAssessments()) {
146			return [];
147			}
148			$pageAssessmentsTable = $this->getTableName($project->getDatabaseName(), 'page_assessments');
			0 ignored issues – show Unused Code introduced 2017-06-30 19:24 UTC by Report Bug Copy Issue Report `$pageAssessmentsTable` is not used, you could remove the assignment. This check looks for variable assignements that are either overwritten by other assignments or where the variable is not used subsequently. $myVar = 'Value'; $higher = false; if (rand(1, 6) > 3) { $higher = true; } else { $higher = false; } Both the `$myVar` assignment in line 1 and the `$higher` assignment in line 2 are dead. The first because `$myVar` is never used and the second because `$higher` is always overwritten for every possible time line. Loading history...
149			$pageIds = implode($pageIds, ',');
150
151			$query = "SELECT pap_project_title AS wikiproject, pa_class AS class, pa_importance AS importance
152			FROM page_assessments
153			LEFT JOIN page_assessments_projects ON pa_project_id = pap_project_id
154			WHERE pa_page_id IN ($pageIds)";
155
156			$conn = $this->getProjectsConnection();
157			return $conn->executeQuery($query)->fetchAll();
158			}
159
160			/**
161			* Get any CheckWiki errors of a single page
162			* @param Page $page
163			* @return array Results from query
164			*/
165			public function getCheckWikiErrors(Page $page)
166			{
167			// Only support mainspace on Labs installations
168			if ($page->getNamespace() !== 0 \|\| !$this->isLabs()) {
169			return [];
170			}
171
172			$sql = "SELECT error, notice, found, name_trans AS name, prio, text_trans AS explanation
173			FROM s51080__checkwiki_p.cw_error a
174			JOIN s51080__checkwiki_p.cw_overview_errors b
175			WHERE a.project = b.project
176			AND a.project = :dbName
177			AND a.title = :title
178			AND a.error = b.id
179			AND a.ok = 0";
180
181			// remove _p if present
182			$dbName = preg_replace('/_p$/', '', $page->getProject()->getDatabaseName());
183
184			// Page title without underscores (str_replace just to be sure)
185			$pageTitle = str_replace('_', ' ', $page->getTitle());
186
187			$resultQuery = $this->getToolsConnection()->prepare($sql);
188			$resultQuery->bindParam(':dbName', $dbName);
189			$resultQuery->bindParam(':title', $pageTitle);
190			$resultQuery->execute();
191
192			return $resultQuery->fetchAll();
193			}
194
195			/**
196			* Get basic wikidata on the page: label and description.
197			* @param Page $page
198			* @return string[] In the format:
199			* [[
200			* 'term' => string such as 'label',
201			* 'term_text' => string (value for 'label'),
202			* ], ... ]
203			*/
204			public function getWikidataInfo(Page $page)
205			{
206			if (empty($page->getWikidataId())) {
207			return [];
208			}
209
210			$wikidataId = ltrim($page->getWikidataId(), 'Q');
211			$lang = $page->getProject()->getLang();
212
213			$sql = "SELECT IF(term_type = 'label', 'label', 'description') AS term, term_text
214			FROM wikidatawiki_p.wb_entity_per_page
215			JOIN wikidatawiki_p.page ON epp_page_id = page_id
216			JOIN wikidatawiki_p.wb_terms ON term_entity_id = epp_entity_id
217			AND term_language = :lang
218			AND term_type IN ('label', 'description')
219			WHERE epp_entity_id = :wikidataId
220
221			UNION
222
223			SELECT pl_title AS term, wb_terms.term_text
224			FROM wikidatawiki_p.pagelinks
225			JOIN wikidatawiki_p.wb_terms ON term_entity_id = SUBSTRING(pl_title, 2)
226			AND term_entity_type = (IF(SUBSTRING(pl_title, 1, 1) = 'Q', 'item', 'property'))
227			AND term_language = :lang
228			AND term_type = 'label'
229			WHERE pl_namespace IN (0, 120)
230			AND pl_from = (
231			SELECT page_id FROM page
232			WHERE page_namespace = 0
233			AND page_title = 'Q:wikidataId'
234			)";
235
236			$resultQuery = $this->getProjectsConnection()->prepare($sql);
237			$resultQuery->bindParam(':lang', $lang);
238			$resultQuery->bindParam(':wikidataId', $wikidataId);
239			$resultQuery->execute();
240
241			return $resultQuery->fetchAll();
242			}
243
244			/**
245			* Get or count all wikidata items for the given page,
246			* not just languages of sister projects
247			* @param Page $page
248			* @param bool $count Set to true to get only a COUNT
249			* @return string[]\|int Records as returend by the DB,
250			* or raw COUNT of the records.
251			*/
252			public function getWikidataItems(Page $page, $count = false)
253			{
254			if (!$page->getWikidataId()) {
255			return $count ? 0 : [];
256			}
257
258			$wikidataId = ltrim($page->getWikidataId(), 'Q');
259
260			$sql = "SELECT " . ($count ? 'COUNT() AS count' : '') . "
261			FROM wikidatawiki_p.wb_items_per_site
262			WHERE ips_item_id = :wikidataId";
263
264			$resultQuery = $this->getProjectsConnection()->prepare($sql);
265			$resultQuery->bindParam(':wikidataId', $wikidataId);
266			$resultQuery->execute();
267
268			$result = $resultQuery->fetchAll();
269
270			return $count ? (int) $result[0]['count'] : $result;
271			}
272
273			/**
274			* Count wikidata items for the given page, not just languages of sister projects
275			* @param Page $page
276			* @return int Number of records.
277			*/
278			public function countWikidataItems(Page $page)
279			{
280			return $this->getWikidataItems($page, true);
281			}
282			}
283

x-tools / xtools

Push — master ( f9a368...59d319 )

PagesRepository::getWikidataInfo() B

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like