Issues in ZendBasedParser.php (master) - Issues in master - kosuha606/html-uni-parser - Measure and Improve Code Quality continuously with Scrutinizer

Issues (8)

src/ZendBasedParser.php (1 issue)

Labels

Bug 1

Severity

<?php
declare(strict_types=1);

namespace kosuha606\HtmlUniParser;

use kosuha606\HtmlUniParser\exceptions\ParserInvalidConfigException;
use PhantomInstaller\PhantomBinary;
filter:
    dependency_paths: ["lib/*"]
use Zend\Dom\Query;

/**
 * Parser helper based on zend query
 * @package app\Parsers
 */
class ZendBasedParser extends BaseObject
{
    /**
     * @var string
     */
    private $url;

    /**
     * @var string
     */
    private $lastUrl;

    /**
     * @var string
     */
    private $htmlBuffer;

    /**
     * @var int
     */
    private $sleepAfterRequest = 0;

    /**
     * @var string
     */
    private $userAgent = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.36";

    /**
     * @param $url
     */
    public function setUrl($url)
    {
        $this->url = $url;
    }

    /**
     * @param int $sleepAfterRequest
     */
    public function setSleepAfterRequest($sleepAfterRequest)
    {
        $this->sleepAfterRequest = $sleepAfterRequest;
    }

    /**
     * @param $search
     * @return mixed
     * @throws ParserInvalidConfigException
     */
    public function query($search)
    {
        return $this->dom()->execute($search);
    }

    /**
     * @return Query
     */
    public function filegetcontentsDom(): Query
    {
        if ($this->getLastUrl() !== $this->getUrl()) {
            $this
                ->setHtmlBuffer('<meta charset="UTF-8" />'.file_get_contents($this->getUrl()))
                ->setLastUrl($this->getUrl())
            ;
            sleep($this->sleepAfterRequest);
        }
        $dom = new Query($this->htmlBuffer);
        return $dom;
    }

    /**
     * @return Query
     */
    public function phantomjsDom(): Query
    {
        if ($this->getLastUrl() !== $this->getUrl()) {
            $bin = PhantomBinary::BIN;
            $command = $bin.' '.__DIR__.'/nodejs/loadspeed.js ' . $this->url;
            $result = shell_exec($command);
            $this
                ->setHtmlBuffer('<meta charset="UTF-8" />'.$result)
                ->setLastUrl($this->getUrl())
            ;
            sleep($this->sleepAfterRequest);
        }
        $dom = new Query($this->htmlBuffer);
        return $dom;
    }

    /**
     * @return Query
     */
    public function wgetDom(): Query
    {
        if ($this->getLastUrl() !== $this->getUrl()) {
            $command = 'wget -qO- '.$this->url.' --no-check-certificate';
            $result = shell_exec($command);
            $this
                ->setHtmlBuffer('<meta charset="UTF-8" />'.$result)
                ->setLastUrl($this->getUrl())
            ;
            sleep($this->sleepAfterRequest);
        }
        $dom = new Query($this->htmlBuffer);
        return $dom;
    }

    /**
     * @return Query
     * @throws ParserInvalidConfigException
     */
    public function dom($encoding = 'UTF-8', $type='curl'): Query
    {
        if ($type==='curl') {
            if (!in_array('curl', get_loaded_extensions())) {
                throw new ParserInvalidConfigException('The curl extension in not loaded in system');
            }
            if ($this->getLastUrl() !== $this->getUrl()) {
                $ch = \curl_init($this->getUrl());
                \curl_setopt($ch, CURLOPT_USERAGENT, $this->userAgent);
                \curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
                \ob_start();
                \ob_implicit_flush();
                \curl_exec($ch);
                \curl_close($ch);
                $htmlBuffer = '<meta charset="UTF-8" />';
                $htmlBuffer .= \ob_get_clean();
                $this->setHtmlBuffer($htmlBuffer);
                $this->setLastUrl($this->getUrl());
                \sleep($this->getSleepAfterRequest());
            }
            $emptyHtml = '<html><head></head><body></body></html>';
            $html = $this->getHtmlBuffer();
            if (!$html) {
                $html = $emptyHtml;
            }
            $dom = new Query($html, $encoding);
            return $dom;
        } else {
            $method = $type.'Dom';
            return $this->$method();
        }
    }

    /**
     * @return string
     */
    public function getRawHtml(): string
    {
        return $this->htmlBuffer ?: '';
    }

    /**
     * @param $html
     * @return ZendBasedParser
     */
    public function setRawHtml($html): self
    {
        $this->htmlBuffer = $html;
        return $this;
    }

    /**
     * @return string
     */
    public function getLastUrl(): string
    {
        return $this->lastUrl ?: '';
    }

    /**
     * @return string
     */
    public function getUrl(): string
    {
        return $this->url;
    }

    /**
     * @return string
     */
    public function getHtmlBuffer(): string
    {
        return $this->htmlBuffer;
    }

    /**
     * @return int
     */
    public function getSleepAfterRequest(): int
    {
        return $this->sleepAfterRequest;
    }

    /**
     * @return string
     */
    public function getUserAgent(): string
    {
        return $this->userAgent;
    }

    /**
     * @param string $lastUrl
     * @return ZendBasedParser
     */
    public function setLastUrl($lastUrl): self
    {
        $this->lastUrl = $lastUrl;
        return $this;
    }

    /**
     * @param string $htmlBuffer
     * @return ZendBasedParser
     */
    public function setHtmlBuffer($htmlBuffer): self
    {
        $this->htmlBuffer = $htmlBuffer;
        return $this;
    }

    /**
     * @param string $userAgent
     * @return ZendBasedParser
     */
    public function setUserAgent($userAgent): self
    {
        $this->userAgent = $userAgent;
        return $this;
    }
}

1			<?php
2			declare(strict_types=1);
3
4			namespace kosuha606\HtmlUniParser;
5
6			use kosuha606\HtmlUniParser\exceptions\ParserInvalidConfigException;
7			use PhantomInstaller\PhantomBinary;
			0 ignored issues – show Bug introduced 2019-07-01 08:30 UTC by Report Bug Copy Issue Report Show Similar Issues like this The type `PhantomInstaller\PhantomBinary` was not found. Maybe you did not declare it correctly or list all dependencies? The issue could also be caused by a filter entry in the build configuration. If the path has been excluded in your configuration, e.g. `excluded_paths: ["lib/"]`, you can move it to the dependency path list as follows: filter: dependency_paths: ["lib/"] For further information see https://scrutinizer-ci.com/docs/tools/php/php-scrutinizer/#list-dependency-paths Loading history...
8			use Zend\Dom\Query;
9
10			/**
11			* Parser helper based on zend query
12			* @package app\Parsers
13			*/
14			class ZendBasedParser extends BaseObject
15			{
16			/**
17			* @var string
18			*/
19			private $url;
20
21			/**
22			* @var string
23			*/
24			private $lastUrl;
25
26			/**
27			* @var string
28			*/
29			private $htmlBuffer;
30
31			/**
32			* @var int
33			*/
34			private $sleepAfterRequest = 0;
35
36			/**
37			* @var string
38			*/
39			private $userAgent = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.36";
40
41			/**
42			* @param $url
43			*/
44			public function setUrl($url)
45			{
46			$this->url = $url;
47			}
48
49			/**
50			* @param int $sleepAfterRequest
51			*/
52			public function setSleepAfterRequest($sleepAfterRequest)
53			{
54			$this->sleepAfterRequest = $sleepAfterRequest;
55			}
56
57			/**
58			* @param $search
59			* @return mixed
60			* @throws ParserInvalidConfigException
61			*/
62			public function query($search)
63			{
64			return $this->dom()->execute($search);
65			}
66
67			/**
68			* @return Query
69			*/
70			public function filegetcontentsDom(): Query
71			{
72			if ($this->getLastUrl() !== $this->getUrl()) {
73			$this
74			->setHtmlBuffer('<meta charset="UTF-8" />'.file_get_contents($this->getUrl()))
75			->setLastUrl($this->getUrl())
76			;
77			sleep($this->sleepAfterRequest);
78			}
79			$dom = new Query($this->htmlBuffer);
80			return $dom;
81			}
82
83			/**
84			* @return Query
85			*/
86			public function phantomjsDom(): Query
87			{
88			if ($this->getLastUrl() !== $this->getUrl()) {
89			$bin = PhantomBinary::BIN;
90			$command = $bin.' '.__DIR__.'/nodejs/loadspeed.js ' . $this->url;
91			$result = shell_exec($command);
92			$this
93			->setHtmlBuffer('<meta charset="UTF-8" />'.$result)
94			->setLastUrl($this->getUrl())
95			;
96			sleep($this->sleepAfterRequest);
97			}
98			$dom = new Query($this->htmlBuffer);
99			return $dom;
100			}
101
102			/**
103			* @return Query
104			*/
105			public function wgetDom(): Query
106			{
107			if ($this->getLastUrl() !== $this->getUrl()) {
108			$command = 'wget -qO- '.$this->url.' --no-check-certificate';
109			$result = shell_exec($command);
110			$this
111			->setHtmlBuffer('<meta charset="UTF-8" />'.$result)
112			->setLastUrl($this->getUrl())
113			;
114			sleep($this->sleepAfterRequest);
115			}
116			$dom = new Query($this->htmlBuffer);
117			return $dom;
118			}
119
120			/**
121			* @return Query
122			* @throws ParserInvalidConfigException
123			*/
124			public function dom($encoding = 'UTF-8', $type='curl'): Query
125			{
126			if ($type==='curl') {
127			if (!in_array('curl', get_loaded_extensions())) {
128			throw new ParserInvalidConfigException('The curl extension in not loaded in system');
129			}
130			if ($this->getLastUrl() !== $this->getUrl()) {
131			$ch = \curl_init($this->getUrl());
132			\curl_setopt($ch, CURLOPT_USERAGENT, $this->userAgent);
133			\curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
134			\ob_start();
135			\ob_implicit_flush();
136			\curl_exec($ch);
137			\curl_close($ch);
138			$htmlBuffer = '<meta charset="UTF-8" />';
139			$htmlBuffer .= \ob_get_clean();
140			$this->setHtmlBuffer($htmlBuffer);
141			$this->setLastUrl($this->getUrl());
142			\sleep($this->getSleepAfterRequest());
143			}
144			$emptyHtml = '<html><head></head><body></body></html>';
145			$html = $this->getHtmlBuffer();
146			if (!$html) {
147			$html = $emptyHtml;
148			}
149			$dom = new Query($html, $encoding);
150			return $dom;
151			} else {
152			$method = $type.'Dom';
153			return $this->$method();
154			}
155			}
156
157			/**
158			* @return string
159			*/
160			public function getRawHtml(): string
161			{
162			return $this->htmlBuffer ?: '';
163			}
164
165			/**
166			* @param $html
167			* @return ZendBasedParser
168			*/
169			public function setRawHtml($html): self
170			{
171			$this->htmlBuffer = $html;
172			return $this;
173			}
174
175			/**
176			* @return string
177			*/
178			public function getLastUrl(): string
179			{
180			return $this->lastUrl ?: '';
181			}
182
183			/**
184			* @return string
185			*/
186			public function getUrl(): string
187			{
188			return $this->url;
189			}
190
191			/**
192			* @return string
193			*/
194			public function getHtmlBuffer(): string
195			{
196			return $this->htmlBuffer;
197			}
198
199			/**
200			* @return int
201			*/
202			public function getSleepAfterRequest(): int
203			{
204			return $this->sleepAfterRequest;
205			}
206
207			/**
208			* @return string
209			*/
210			public function getUserAgent(): string
211			{
212			return $this->userAgent;
213			}
214
215			/**
216			* @param string $lastUrl
217			* @return ZendBasedParser
218			*/
219			public function setLastUrl($lastUrl): self
220			{
221			$this->lastUrl = $lastUrl;
222			return $this;
223			}
224
225			/**
226			* @param string $htmlBuffer
227			* @return ZendBasedParser
228			*/
229			public function setHtmlBuffer($htmlBuffer): self
230			{
231			$this->htmlBuffer = $htmlBuffer;
232			return $this;
233			}
234
235			/**
236			* @param string $userAgent
237			* @return ZendBasedParser
238			*/
239			public function setUserAgent($userAgent): self
240			{
241			$this->userAgent = $userAgent;
242			return $this;
243			}
244			}

kosuha606 / html-uni-parser

Issues (8)

src/ZendBasedParser.php (1 issue)

Labels

Severity

Introduced By

Duplication Side-by-Side

Filter issues like