For a survey I am trying to grab google search result from my php page. I grab six results then i click next button to get next page records.but after 9 pages e.g. 64 results it gives following error:
stdClass Object
(
[responseData] =>
[responseDetails] => out of range start
[responseStatus] => 400
)
I just want as much data as possible. i dont mind if it is google search engine or any other search engine. but for accurate result of survey I want large number of resultset. Can anybody knows how can i do that ?
Is it possible to grab results through cron ? Is there any other way ?
ATTENTION
Google tries to prevent scraping and so servers will be blocked and requests will be dropped when they suspect scraping. So you can use this if you occassionally need to get some google results. Check google-scraper.squabbel.com for a proxy based scraper and more information on googles blocking mechanism. Its also against their policy and thus illigal.
The google api will not allow for more then 64 results, so if you need more you need to scrape the site yourself. As it was a fun project, I have created a class to do this for you.
It requires the free PHP Simple HTML DOM Parser so you need download this code aswell.
it will output an array like
array(100) {
[0]=>
array(3) {
["title"]=>
string(67) "Online Tests - Online aptitude tests for interview, competitive ..."
["href"]=>
string(36) "http://www.indiabix.com/online-test/"
["description"]=>
string(168) "Online aptitude tests for competitive examination, entrance examination and
campus interview. Take various tests and find out how much you score before
you ... "
}
[1]=>
array(3) {
["title"]=>
string(37) "Test your English - Cambridge English"
["href"]=>
string(50) "http://www.cambridgeenglish.org/test-your-english/"
["description"]=>
string(179) "Test Your English. This is a quick, free online test. It will tell you which Cambridge
English exam may be best for you. Click 'Begin Test' and answer each of the ... "
}
//removed for better visibility
}
How to use:
//start the scraper for google.com (english results)
$gs = new GoogleScraper();
//start the scraper for google.nl (dutch results)
//$gs = new GoogleScraper('https://www.google.nl');
//set your search query
$gs->SearchQuery('online exams');
//start loading the pages. You can enter any integer above 0
$gs->LoadPages(10);
//dump the results, but its just an array so you can also do other things with it.
echo '<pre>';
var_dump($gs->GetResults());
echo '</pre>';
?>
And then the GoogleScraper.php
<?php
require_once('simple_html_dom.php');
class GoogleScraper
{
private $_results;
private $_baseUrl;
private $_searchQuery;
private $_resultsPerPage;
/**
* constructor
* I use the constructor to set all the defaults to keep it all in one place
*/
final public function __construct($baseUrl='')
{
$this->_results = array();
$this->_resultsPerPage = 100;
if (empty($baseUrl)) {
$this->_baseUrl = 'https://www.google.com';
} else {
$this->_baseUrl = $baseUrl;
}
}
/**
* cleanup
*/
final public function __destruct()
{
unset($this->_results);
unset($this->_baseUrl);
unset($this->_searchQuery);
}
/**
* Set the query
*/
final public function SearchQuery($searchQuery)
{
if (!(is_string($searchQuery) || is_numeric($searchQuery)))
{
throw new Exception('Invalid query type');
}
$this->_searchQuery = $searchQuery;
}
/**
* Set the number of results per page
*/
final public function ResultsPerPage($resultsPerPage)
{
if (!is_int($resultsPerPage) || $resultsPerPage<10 || $resultsPerPage>100)
{
throw new Exception('Results per page must be value between 10 and 100');
}
$this->_resultsPerPage = $resultsPerPage;
}
/**
* Get the result
*/
final public function GetResults()
{
return $this->_results;
}
/**
* Scrape the search results
*/
final public function LoadPages($pages=1)
{
if (!is_int($pages) || $pages<1)
{
throw new Exception('Invalid number of pages');
}
if (empty($this->_searchQuery))
{
throw new Exception('Missing search query');
}
$url = $this->_baseUrl . '/search?num='.$this->_resultsPerPage.'&q=' . urlencode($this->_searchQuery);
$currentPage = 1;
while($pages--) {
if ($content = $this->LoadUrl($url)) {
/*
Load content in to simple html dom
*/
$html = new simple_html_dom();
$html->load($content);
/*
Find and handle search results
*/
$items = $html->find('div#ires li');
foreach($items as $item) {
/*
Only normal search results have this container. Special results like found images or news dont have it.
*/
$check = $item->find('div.s');
if (count($check)!=1) {
continue;
}
$head = $item->find('h3.r a', 0);
$result['title'] = $head->plaintext;
/*
If we dont have a title, there is no point in continuing
*/
if (empty($result['title'])) {
continue;
}
$result['href'] = $head->href;
/*
Check if we can parse the URL for the actual url
*/
if (!empty($result['href'])) {
$qs = explode('?', $result['href']);
if (!empty($qs[1])) {
parse_str($qs[1], $querystring);
if (!empty($querystring['q'])) {
$result['href'] = $querystring['q'];
}
}
}
/*
Try to find the description
*/
$info = $item->find('span.st', 0);
$result['description'] = $info->plaintext;
/*
Add the results to the total
*/
$this->_results[] = $result;
}
/*
Find next page
*/
$url = $this->_baseUrl . '/search?num='.$this->_resultsPerPage.'&q=' . urlencode($this->_searchQuery) . '$start=' . ($currentPage*$this->_resultsPerPage);
} else {
throw new Exception('Failed to load page');
}
$currentPage++;
}
}
/**
* Load the url
*/
final private function LoadUrl($url)
{
if (!is_string($url))
{
throw new Exception('Invalid url');
}
$options['http'] = array(
'user_agent' => "GoogleScraper",
'timeout' => 0.5
);
$context = stream_context_create($options);
$content = file_get_contents($url, null, $context);
if (!empty($http_response_header))
{
return (substr_count($http_response_header[0], ' 200 OK')>0) ? $content : false;
}
return false;
}
}
?>
Check this PHP Fiddle to see it in action. Because this could be used quite often from this server, there is a chance for 503 errors from google.