/**
* {@inheritDoc}
*
* @param string $page the image represented as a character string
* @param string $url the url where the image was downloaded from
* @return array summary information including a thumbnail and a
* description (where the description is just the url)
*/
function process($page, $url)
{
if (is_string($page)) {
$image = $this->imagecreatefrombmp($page);
$thumb_string = self::createThumb($image);
$summary[self::TITLE] = "";
$summary[self::DESCRIPTION] = "Image of " . UrlParser::getDocumentFilename($url);
$summary[self::LINKS] = array();
$summary[self::PAGE] = "<html><body><div><img src='data:image/bmp;base64," . base64_encode($page) . "' alt='" . $summary[self::DESCRIPTION] . "' /></div></body></html>";
$summary[self::THUMB] = 'data:image/jpeg;base64,' . base64_encode($thumb_string);
}
return $summary;
}
/**
* Prepare an array of up to NUM_MULTI_CURL_PAGES' worth of sites to be
* downloaded in one go using the to_crawl array. Delete these sites
* from the to_crawl array.
*
* @return array sites which are ready to be downloaded
*/
function getFetchSites()
{
$web_archive = $this->web_archive;
$start_time = microtime();
$seeds = array();
$delete_indices = array();
$num_items = count($this->to_crawl);
if ($num_items > 0) {
$crawl_source =& $this->to_crawl;
$to_crawl_flag = true;
} else {
crawlLog("...Trying to crawl sites which failed the first time");
$num_items = count($this->to_crawl_again);
$crawl_source =& $this->to_crawl_again;
$to_crawl_flag = false;
}
reset($crawl_source);
if ($num_items > NUM_MULTI_CURL_PAGES) {
$num_items = NUM_MULTI_CURL_PAGES;
}
//DNS lookups take longer so try to get fewer in one go
$num_ip_lookups = max($num_items / 3, 2);
$i = 0;
$ip_lookup_cnt = 0;
$site_pair = each($crawl_source);
while ($site_pair !== false && $i < $num_items && $ip_lookup_cnt < $num_ip_lookups) {
$delete_indices[] = $site_pair['key'];
if ($site_pair['value'][0] != self::DUMMY) {
$host = UrlParser::getHost($site_pair['value'][0]);
if (!strpos($site_pair['value'][0], "###")) {
$ip_lookup_cnt++;
}
// only download if host doesn't seem congested
if (!isset($this->hosts_with_errors[$host]) || $this->hosts_with_errors[$host] < DOWNLOAD_ERROR_THRESHOLD) {
$url_to_check = $site_pair['value'][0];
$extension = UrlParser::getDocumentType($url_to_check);
$repository_indicator = FetchGitRepositoryUrls::checkForRepository($extension);
if ($repository_indicator == self::REPOSITORY_GIT) {
$git_internal_urls = FetchGitRepositoryUrls::setGitRepositoryUrl($url_to_check, $i, $seeds, $repository_indicator, $site_pair, $this->total_git_urls, $this->all_git_urls);
$i = $git_internal_urls['position'];
$git_url_index = $git_internal_urls['index'];
$seeds = $git_internal_urls['seeds'];
$repository_indicator = $git_internal_urls['indicator'];
$this->total_git_urls = $git_internal_urls['count'];
$this->all_git_urls = $git_internal_urls['all'];
} else {
$seeds[$i][self::URL] = $site_pair['value'][0];
$seeds[$i][self::WEIGHT] = $site_pair['value'][1];
$seeds[$i][self::CRAWL_DELAY] = $site_pair['value'][2];
}
/*
Crawl delay is only used in scheduling on the queue_server
on the fetcher, we only use crawl-delay to determine
if we will give a page a second try if it doesn't
download the first time
*/
if (UrlParser::getDocumentFilename($seeds[$i][self::URL]) . "." . UrlParser::getDocumentType($seeds[$i][self::URL]) == "robots.txt") {
$seeds[$i][self::ROBOT_PATHS] = array();
}
$i++;
}
} else {
break;
}
$site_pair = each($crawl_source);
}
//end while
foreach ($delete_indices as $delete_index) {
$git_set = false;
if ($to_crawl_flag == true) {
$extension = UrlParser::getDocumentType($this->to_crawl[$delete_index][0]);
$repository_type = FetchGitRepositoryUrls::checkForRepository($extension);
if ($repository_type != self::REPOSITORY_GIT) {
unset($this->to_crawl[$delete_index]);
}
} else {
$extension = UrlParser::getDocumentType($this->to_crawl_again[$delete_index][0]);
$repository_type = FetchGitRepositoryUrls::checkForRepository($extension);
unset($this->to_crawl_again[$delete_index]);
}
if ($repository_type == self::REPOSITORY_GIT) {
if (!$git_set) {
$next_url_start = $url_to_check . self::GIT_URL_CONTINUE . $git_url_index;
$git_set = true;
$this->to_crawl[$delete_index][0] = $next_url_start;
}
if ($repository_indicator == self::INDICATOR_NONE) {
unset($this->to_crawl[$delete_index]);
}
}
}
crawlLog("Fetch url list to download time " . changeInMicrotime($start_time));
return $seeds;
}
/**
* Make multi_curl requests for an array of sites with urls or onion urls
*
* @param array $sites an array containing urls of pages to request
* @param bool $timer flag, true means print timing statistics to log
* @param int $page_range_request maximum number of bytes to download/page
* 0 means download all
* @param string $temp_dir folder to store temporary ip header info
* @param string $key the component of $sites[$i] that has the value of
* a url to get defaults to URL
* @param string $value component of $sites[$i] in which to store the
* page that was gotten
* @param bool $minimal if true do a faster request of pages by not
* doing things like extract HTTP headers sent, etcs
* @param array $post_data data to be POST'd to each site
* @param bool $follow whether to follow redirects or not
* @param string $tor_proxy url of a proxy that knows how to download
* .onion urls
* @param array $proxy_servers if not array(), then an array of proxy
* server to use rather than to directly download web pages from
* the current machine
*
* @return array an updated array with the contents of those pages
*/
static function getPages($sites, $timer = false, $page_range_request = PAGE_RANGE_REQUEST, $temp_dir = NULL, $key = CrawlConstants::URL, $value = CrawlConstants::PAGE, $minimal = false, $post_data = NULL, $follow = false, $tor_proxy = "", $proxy_servers = array())
{
$agent_handler = curl_multi_init();
$active = NULL;
$start_time = microtime();
if (!$minimal && $temp_dir == NULL) {
$temp_dir = CRAWL_DIR . "/temp";
if (!file_exists($temp_dir)) {
mkdir($temp_dir);
}
}
//Set-up requests
$num_sites = count($sites);
for ($i = 0; $i < $num_sites; $i++) {
$is_gopher = false;
$sites[$i][CrawlConstants::IS_GOPHER_URL] = $is_gopher;
if (isset($sites[$i][$key])) {
list($sites[$i][$key], $url, $headers) = self::prepareUrlHeaders($sites[$i][$key], $minimal, $proxy_servers);
if ($headers == "gopher") {
$is_gopher = true;
$sites[$i][CrawlConstants::IS_GOPHER_URL] = $is_gopher;
$headers = array();
}
$sites[$i][0] = curl_init();
if (!$minimal) {
$ip_holder[$i] = fopen("{$temp_dir}/tmp{$i}.txt", 'w+');
curl_setopt($sites[$i][0], CURLOPT_STDERR, $ip_holder[$i]);
curl_setopt($sites[$i][0], CURLOPT_VERBOSE, true);
}
curl_setopt($sites[$i][0], CURLOPT_USERAGENT, USER_AGENT);
curl_setopt($sites[$i][0], CURLOPT_IPRESOLVE, CURL_IPRESOLVE_WHATEVER);
curl_setopt($sites[$i][0], CURLOPT_URL, $url);
if (strcmp(substr($url, -10), "robots.txt") == 0) {
$sites[$i]['ROBOT'] = true;
$follow = true;
/*wikipedia redirects their robot page. grr
want to force this for robots pages
*/
}
curl_setopt($sites[$i][0], CURLOPT_FOLLOWLOCATION, $follow);
curl_setopt($sites[$i][0], CURLOPT_SSL_VERIFYHOST, 0);
curl_setopt($sites[$i][0], CURLOPT_AUTOREFERER, true);
curl_setopt($sites[$i][0], CURLOPT_RETURNTRANSFER, true);
curl_setopt($sites[$i][0], CURLOPT_CONNECTTIMEOUT, PAGE_TIMEOUT);
curl_setopt($sites[$i][0], CURLOPT_TIMEOUT, PAGE_TIMEOUT);
if (stripos($url, '.onion') !== false && $tor_proxy != "") {
curl_setopt($sites[$i][0], CURLOPT_PROXY, $tor_proxy);
//CURLPROXY_SOCKS5_HOSTNAME = 7
curl_setopt($sites[$i][0], CURLOPT_PROXYTYPE, 7);
if ($timer) {
crawlLog("Using Tor proxy for {$url}..");
}
} else {
if ($proxy_servers != array() && !$is_gopher) {
$select_proxy = rand(0, count($proxy_servers) - 1);
$proxy_server = $proxy_servers[$select_proxy];
$proxy_parts = explode(":", $proxy_server);
$proxy_ip = $proxy_parts[0];
if (!isset($proxy_parts[2]) || strtolower($proxy_parts[2]) == 'http') {
$proxy_type = CURLPROXY_HTTP;
} else {
if (strtolower($proxy_parts[2]) == 'socks5') {
$proxy_type = CURLPROXY_SOCKS5;
} else {
$proxy_type = $proxy_parts[2];
}
}
if (isset($proxy_parts[1])) {
$proxy_port = $proxy_parts[1];
} else {
$proxy_port = "80";
}
curl_setopt($sites[$i][0], CURLOPT_PROXY, "{$proxy_ip}:{$proxy_port}");
curl_setopt($sites[$i][0], CURLOPT_PROXYTYPE, $proxy_type);
if ($timer) {
crawlLog("Selecting proxy {$select_proxy} for {$url}");
}
}
}
if (!$minimal) {
curl_setopt($sites[$i][0], CURLOPT_HEADER, true);
}
//make lighttpd happier
if (!$is_gopher) {
curl_setopt($sites[$i][0], CURLOPT_HTTPHEADER, $headers);
}
curl_setopt($sites[$i][0], CURLOPT_ENCODING, "");
// ^ need to set for sites like att that use gzip
if ($page_range_request > 0) {
curl_setopt($sites[$i][0], CURLOPT_RANGE, "0-" . $page_range_request);
}
if ($post_data != NULL) {
curl_setopt($sites[$i][0], CURLOPT_POST, true);
curl_setopt($sites[$i][0], CURLOPT_POSTFIELDS, $post_data[$i]);
}
curl_multi_add_handle($agent_handler, $sites[$i][0]);
}
}
if ($timer) {
crawlLog(" Init Get Pages " . changeInMicrotime($start_time));
}
$start_time = microtime();
$start = time();
//Wait for responses
$running = NULL;
$memory_limit = metricToInt(ini_get("memory_limit")) * 0.7;
do {
$mrc = curl_multi_exec($agent_handler, $running);
$ready = curl_multi_select($agent_handler, 0.005);
} while (memory_get_usage() < $memory_limit && time() - $start < PAGE_TIMEOUT && $running > 0);
if (time() - $start > PAGE_TIMEOUT && $timer) {
crawlLog(" TIMED OUT!!!");
}
if ($timer) {
crawlLog(" Page Request time " . changeInMicrotime($start_time));
}
$start_time = microtime();
//Process returned pages
for ($i = 0; $i < $num_sites; $i++) {
if ($timer) {
crawlTimeoutLog("fetch_url initial processing of page %s of %s", $i, $num_sites);
}
if (!$minimal && isset($ip_holder[$i])) {
rewind($ip_holder[$i]);
$header = fread($ip_holder[$i], 8192);
$ip_addresses = self::getCurlIp($header);
fclose($ip_holder[$i]);
}
$is_gopher = false;
if (isset($sites[$i][0]) && $sites[$i][0]) {
// Get Data and Message Code
$content = @curl_multi_getcontent($sites[$i][0]);
$is_gopher = $sites[$i][CrawlConstants::IS_GOPHER_URL];
/*
If the Transfer-encoding was chunked then the Range header
we sent was ignored. So we manually truncate the data
here
*/
if ($page_range_request > 0) {
$content = substr($content, 0, $page_range_request);
}
if (isset($content) && !$minimal && !$is_gopher) {
$site = self::parseHeaderPage($content, $value);
$sites[$i] = array_merge($sites[$i], $site);
if (isset($header)) {
$header = substr($header, 0, strpos($header, "\r\n\r\n") + 4);
} else {
$header = "";
}
$sites[$i][CrawlConstants::HEADER] = $header . $sites[$i][CrawlConstants::HEADER];
unset($header);
} else {
if (isset($content) && !$minimal && $is_gopher) {
$sites[$i][CrawlConstants::HEADER] = $header;
$sites[$i][$value] = $content;
unset($header);
} else {
$sites[$i][$value] = $content;
}
}
if (!$minimal) {
$sites[$i][self::SIZE] = @curl_getinfo($sites[$i][0], CURLINFO_SIZE_DOWNLOAD);
$sites[$i][self::DNS_TIME] = @curl_getinfo($sites[$i][0], CURLINFO_NAMELOOKUP_TIME);
$sites[$i][self::TOTAL_TIME] = @curl_getinfo($sites[$i][0], CURLINFO_TOTAL_TIME);
$sites[$i][self::HTTP_CODE] = curl_getinfo($sites[$i][0], CURLINFO_HTTP_CODE);
if (!$sites[$i][self::HTTP_CODE] && !$is_gopher) {
$sites[$i][self::HTTP_CODE] = curl_error($sites[$i][0]);
} else {
$sites[$i][self::HTTP_CODE] = 200;
}
if ($ip_addresses) {
$sites[$i][self::IP_ADDRESSES] = $ip_addresses;
} else {
$sites[$i][self::IP_ADDRESSES] = array("0.0.0.0");
}
//Get Time, Mime type and Character encoding
$sites[$i][self::TIMESTAMP] = time();
if ($is_gopher) {
$path = UrlParser::getPath($sites[$i][self::URL]);
$filename = UrlParser::getDocumentFilename($sites[$i][self::URL]);
if (isset($path[1])) {
$gopher_type = $path[1];
} else {
$gopher_type = 1;
}
if ($gopher_type == 1) {
$sites[$i][self::TYPE] = "text/gopher";
} else {
if (in_array($gopher_type, array(0, 3, 6))) {
$sites[$i][self::TYPE] = "text/plain";
if ($gopher_type == 6) {
$sites[$i][$value] = convert_uudecode($content);
}
} else {
if ($gopher_type == 'h') {
$sites[$i][self::TYPE] = "text/html";
} else {
if ($gopher_type == 'g') {
$sites[$i][self::TYPE] = "image/gif";
}
}
}
}
$path_info = pathinfo($filename);
if (!isset($sites[$i][self::TYPE]) && isset($path_info['extension'])) {
$sites[$i][self::TYPE] = UrlParser::guessMimeTypeFromFileName($filename);
} else {
if (!isset($sites[$i][self::TYPE])) {
$sites[$i][self::TYPE] = "unknown";
}
}
} else {
$type_parts = explode(";", curl_getinfo($sites[$i][0], CURLINFO_CONTENT_TYPE));
$sites[$i][self::TYPE] = strtolower(trim($type_parts[0]));
}
}
//curl_multi_remove_handle($agent_handler, $sites[$i][0]);
curl_close($sites[$i][0]);
if (isset($sites[$i]['ROBOT']) && $sites[$i]['ROBOT']) {
if (isset($sites[$i][self::TYPE]) && $sites[$i][self::TYPE] != "text/plain" && isset($sites[$i][CrawlConstants::LOCATION]) && count($site[CrawlConstants::LOCATION]) > 0) {
$sites[$i][self::TYPE] = "text/plain";
$sites[$i][self::HTTP_CODE] = "200";
$tmp = wordwrap($sites[$i][$value], 80);
$tmp_parts = explode("\n", $tmp);
$tmp = "# Suspect server misconfiguration\n";
$tmp .= "# Assume shouldn't crawl this site.\n";
$tmp .= "# Pretending got following robots.txt.\n";
$tmp .= "User-agent: *\n";
$tmp .= "Disallow: /\n";
$tmp .= "# Original error code: " . $sites[$i][self::HTTP_CODE] . "\n";
$tmp .= "# Original content:\n";
foreach ($tmp_parts as $part) {
$tmp = "#" . $part . "\n";
}
$sites[$i][$value] = $tmp;
$sites[$i][self::HTTP_CODE] = "200";
unset($site[CrawlConstants::LOCATION]);
}
}
}
//end big if
}
//end for
if ($timer) {
crawlLog(" Get Page Content time " . changeInMicrotime($start_time));
}
curl_multi_close($agent_handler);
return $sites;
}
/**
* Used to recompute the dictionary of an index archive -- either from
* scratch using the index shard data or just using the current dictionary
* but merging the tiers into one tier
*
* @param string $path file path to dictionary of an IndexArchiveBundle
* @param int $max_tier tier up to which the dictionary tiers should be
* merge (typically a value greater than the max_tier of the
* dictionary)
*/
function reindexIndexArchive($path, $max_tier = -1)
{
if ($this->getArchiveKind($path) != "IndexArchiveBundle") {
echo "\n{$path} ...\n" . " is not an IndexArchiveBundle so cannot be re-indexed\n\n";
exit;
}
$shards = glob($path . "/posting_doc_shards/index*");
if (is_array($shards)) {
if ($max_tier == -1) {
$dbms_manager = DBMS . "Manager";
$db = new $dbms_manager();
$db->unlinkRecursive($path . "/dictionary", false);
IndexDictionary::makePrefixLetters($path . "/dictionary");
}
$dictionary = new IndexDictionary($path . "/dictionary");
if ($max_tier == -1) {
$max_generation = 0;
foreach ($shards as $shard_name) {
$file_name = UrlParser::getDocumentFilename($shard_name);
$generation = (int) substr($file_name, strlen("index"));
$max_generation = max($max_generation, $generation);
}
for ($i = 0; $i < $max_generation + 1; $i++) {
$shard_name = $path . "/posting_doc_shards/index{$i}";
echo "\nShard {$i}\n";
$shard = new IndexShard($shard_name, $i, NUM_DOCS_PER_GENERATION, true);
$dictionary->addShardDictionary($shard);
}
$max_tier = $dictionary->max_tier;
}
echo "\nFinal Merge Tiers\n";
$dictionary->mergeAllTiers(NULL, $max_tier);
$db->setWorldPermissionsRecursive($path . "/dictionary");
echo "\nReindex complete!!\n";
} else {
echo "\n{$path} ...\n" . " does not contain posting shards so cannot be re-indexed\n\n";
}
}
/**
* Gets the resource $_REQUEST['n'] from APP_DIR/$_REQUEST['f'] or
* CRAWL_DIR/$_REQUEST['f'] after cleaning
*/
function get()
{
if (!isset($_REQUEST['n']) || !isset($_REQUEST['f'])) {
return;
}
$name = $this->clean($_REQUEST['n'], "string");
if (in_array($_REQUEST['f'], array("css", "scripts", "resources"))) {
/* notice in this case we don't check if request come from a
legitimate source but we do try to restrict it to being
a file (not a folder) in the above array
*/
$base_dir = $this->getBaseFolder();
if (!$base_dir) {
header('HTTP/1.1 401 Unauthorized');
echo "<html><head><title>401 Unauthorized</title></head>" . "<body><p>401 Unauthorized</p></body></html>";
return;
}
$type = UrlParser::getDocumentType($name);
$name = UrlParser::getDocumentFilename($name);
$name = $type != "" ? "{$name}.{$type}" : $name;
if (isset($_REQUEST['t'])) {
$name .= ".jpg";
}
} else {
if (in_array($_REQUEST['f'], array("cache"))) {
/* perform check since these request should come from a known
machine
*/
if (!$this->checkRequest()) {
return;
}
$folder = $_REQUEST['f'];
$base_dir = CRAWL_DIR . "/{$folder}";
} else {
return;
}
}
if (isset($_REQUEST['o']) && isset($_REQUEST['l'])) {
$offset = $this->clean($_REQUEST['o'], "int");
$limit = $this->clean($_REQUEST['l'], "int");
}
$path = "{$base_dir}/{$name}";
if (file_exists($path)) {
$mime_type = mimeType($path);
$size = filesize($path);
$start = 0;
$end = $size - 1;
header("Content-type: {$mime_type}");
header("Accept-Ranges: bytes");
if (isset($_SERVER['HTTP_RANGE'])) {
$this->serveRangeRequest($path, $size, $start, $end);
return;
}
header("Content-Length: " . $size);
header("Content-Range: bytes {$start}-{$end}/{$size}");
if (isset($offset) && isset($limit)) {
echo file_get_contents($path, false, NULL, $offset, $limit);
} else {
readfile($path);
}
} else {
header("Location:./error.php");
}
}
