/**
* Sends an email (much like PHP's mail command, but not requiring
* a configured smtp server on the current machine)
*
* @param string $subject subject line of the email
* @param string $from sender email address
* @param string $to recipient email address
* @param string $message message body for the email
*/
function send($subject, $from, $to, $message)
{
$start_time = microtime();
if ($from == "") {
$from = $this->sender_email;
}
$eol = self::EOL;
if (USE_MAIL_PHP) {
$header = "From: " . $from . $eol;
mail($to, $subject, $message, $header);
return;
}
$this->messages = "";
$mail = "Date: " . date(DATE_RFC822) . $eol;
$mail .= "Subject: " . $subject . $eol;
$mail .= "From: " . $from . $eol;
$mail .= "To: " . $to . $eol;
$mail .= $eol . $eol . $message . $eol . ".";
$commands = array("MAIL FROM: <{$from}>" => self::OKAY, "RCPT TO: <{$to}>" => self::OKAY, "DATA" => self::START_INPUT, $mail => self::OKAY);
if ($this->startSession()) {
foreach ($commands as $command => $good_response) {
$response = $this->smtpCommand($command);
if ($response != $good_response) {
$this->messages .= "{$command} failed!! {$response} {$good_response}\n";
break;
}
}
$this->endSession();
}
if (QUERY_STATISTICS) {
$current_messages = AnalyticsManager::get("MAIL_MESSAGES");
if (!$current_messages) {
$current_messages = array();
}
$total_time = AnalyticsManager::get("MAIL_TOTAL_TIME");
if (!$total_time) {
$total_time = 0;
}
$elapsed_time = changeInMicrotime($start_time);
$total_time += $elapsed_time;
$current_messages[] = array("QUERY" => "<p>Send Mail</p>" . "<pre>" . wordwrap($this->messages, 60, "\n", true) . "</pre>", "ELAPSED_TIME" => $elapsed_time);
AnalyticsManager::set("MAIL_MESSAGES", $current_messages);
AnalyticsManager::set("MAIL_TOTAL_TIME", $total_time);
}
}
/**
* Gets doc summaries of documents containing given words and meeting the
* additional provided criteria
* @param array $word_structs an array of word_structs. Here a word_struct
* is an associative array with at least the following fields
* KEYS -- an array of word keys
* QUOTE_POSITIONS -- an array of positions of words that appeared in
* quotes (so need to be matched exactly)
* DISALLOW_PHRASES -- an array of words the document must not contain
* WEIGHT -- a weight to multiple scores returned from this iterator by
* INDEX_NAME -- an index timestamp to get results from
* @param int $limit number of first document in order to return
* @param int $num number of documents to return summaries of
* @param array& $filter an array of hashes of domains to filter from
* results
* @param bool $use_cache_if_allowed if true and USE_CACHE is true then
* an attempt will be made to look up the results in either
* the file cache or memcache. Otherwise, items will be recomputed
* and then potentially restored in cache
* @param int $raw ($raw == 0) normal grouping, ($raw > 0)
* no grouping done on data. if ($raw == 1) no lookups of summaries
* done
* @param array $queue_servers a list of urls of yioop machines which might
* be used during lookup
* @param string $original_query if set, the original query that corresponds
* to $word_structs
* @param string $save_timestamp_name if this timestamp is not empty, then
* save iterate position, so can resume on future queries that make
* use of the timestamp. If used then $limit ignored and get next $num
* docs after $save_timestamp 's previous iterate position.
* @param bool $limit_news if true the number of media:news items to
* allow in search results is limited to WordIterator::LIMIT_NEWS_COUNT
*
* @return array document summaries
*/
function getSummariesByHash($word_structs, $limit, $num, &$filter, $use_cache_if_allowed = true, $raw = 0, $queue_servers = array(), $original_query = "", $save_timestamp_name = "", $limit_news = true)
{
global $CACHE;
$indent = " ";
$in2 = $indent . $indent;
$in3 = $in2 . $indent;
$in4 = $in2 . $in2;
if (QUERY_STATISTICS) {
$lookup_time = microtime();
}
$use_proximity = false;
$time = time();
if (count($word_structs) > 1 || isset($word_structs[0]["KEYS"]) && count($word_structs[0]["KEYS"]) > 1 || $word_structs == array() && substr_count($original_query, " ") > 1) {
$use_proximity = true;
}
if (!isset($filter['time'])) {
$filter['time'] = 0;
}
$filter_time = $filter['time'];
unset($filter['time']);
//iterators don't expect time field
$pages = array();
$generation = 0;
$to_retrieve = ceil(($limit + $num) / self::NUM_CACHE_PAGES) * self::NUM_CACHE_PAGES;
$start_slice = floor($limit / self::NUM_CACHE_PAGES) * self::NUM_CACHE_PAGES;
if ($save_timestamp_name != "") {
$to_retrieve = $num;
$limit = 0;
$start_slice = 0;
}
if (USE_CACHE && $save_timestamp_name == "") {
$mem_tmp = serialize($raw) . serialize($word_structs) . $original_query . $this->index_name;
$summary_hash = crawlHash($mem_tmp . ":" . $limit . ":" . $num);
if ($use_cache_if_allowed) {
$cache_success = true;
$results = $CACHE->get($summary_hash);
if (!isset($results['TIME']) || $filter_time > $results['TIME']) {
//if filter has changed since cached, then invalidate cache
$results = false;
}
if (isset($results['TIME'])) {
$cached_time = $time - $results['TIME'];
} else {
$cached_time = $time;
}
if ($cached_time > MAX_QUERY_CACHE_TIME) {
$results = false;
}
if (isset($results['PAGES'])) {
$close_prefix = WORK_DIRECTORY . "/schedules/" . self::index_closed_name;
$has_changeable_results = false;
$seen_times = array();
foreach ($results['PAGES'] as $page) {
if (!isset($page[self::CRAWL_TIME]) || in_array($page[self::CRAWL_TIME], $seen_times)) {
continue;
}
$seen_times[] = $page[self::CRAWL_TIME];
$current_closed = $close_prefix . $page[self::CRAWL_TIME] . ".txt";
if (!file_exists($current_closed)) {
//either feed result or from active crawl
$has_changeable_results = true;
break;
}
}
if ($has_changeable_results) {
if ($cached_time > MIN_QUERY_CACHE_TIME) {
$results = false;
}
}
}
if (QUERY_STATISTICS) {
$this->query_info['QUERY'] .= "{$in2}<b>Cache Lookup Time</b>: " . changeInMicrotime($lookup_time) . "<br />";
}
if ($results !== false) {
return $results;
}
}
}
$old_to_retrieve = $to_retrieve;
$query_iterator = $this->getQueryIterator($word_structs, $filter, $raw, $to_retrieve, $queue_servers, $original_query, $save_timestamp_name, $limit_news);
$num_retrieved = 0;
$pages = array();
if (is_object($query_iterator)) {
while ($num_retrieved < $to_retrieve && is_array($next_docs = $query_iterator->nextDocsWithWord())) {
$pages += $next_docs;
$num_retrieved = count($pages);
}
}
if ($save_timestamp_name != "" && ($queue_servers == array() || $this->isSingleLocalhost($queue_servers))) {
// used for archive crawls of crawl mixes
$save_file = CRAWL_DIR . '/schedules/' . self::save_point . $save_timestamp_name . ".txt";
$iterators = $query_iterator->save_iterators;
$cnt_iterators = count($iterators);
$save_point = array();
for ($i = 0; $i < $cnt_iterators; $i++) {
$save_point[$i] = $iterators[$i]->currentGenDocOffsetWithWord();
}
$results["SAVE_POINT"] = $save_point;
file_put_contents($save_file, serialize($save_point));
$this->db->setWorldPermissionsRecursive($save_file);
}
$pages = array_values($pages);
$result_count = count($pages);
$sort_time = 0;
if ($raw == 0) {
// initialize scores
$sort_start = microtime();
$max_user_ranks = 0;
for ($i = 0; $i < $result_count; $i++) {
$pages[$i]["OUT_SCORE"] = 0;
if (isset($pages[$i][self::USER_RANKS])) {
$j = count($pages[$i][self::USER_RANKS]);
if ($max_user_ranks < $j) {
$max_user_ranks = $j;
}
}
}
if ($max_user_ranks > 0) {
for ($i = 0; $i < $result_count; $i++) {
for ($j = 0; $j < $max_user_ranks; $j++) {
if (isset($pages[$i][self::USER_RANKS][$j])) {
$pages[$i]["USCORE{$j}"] = $pages[$i][self::USER_RANKS][$j];
} else {
$pages[$i]["USCORE{$j}"] = 0;
}
}
}
}
$subscore_fields = array(self::DOC_RANK, self::RELEVANCE);
if ($use_proximity) {
$subscore_fields[] = self::PROXIMITY;
}
if ($max_user_ranks > 0) {
for ($j = 0; $j < $max_user_ranks; $j++) {
$subscore_fields[] = "USCORE{$j}";
}
}
$num_fields = count($subscore_fields);
// Compute Reciprocal Rank Fusion Score
$alpha = 600 / $num_fields;
if (isset($pages[0])) {
foreach ($subscore_fields as $field) {
orderCallback($pages[0], $pages[0], $field);
usort($pages, "orderCallback");
$score = 0;
for ($i = 0; $i < $result_count; $i++) {
if ($i > 0) {
if ($pages[$i - 1][$field] != $pages[$i][$field]) {
$score++;
}
}
$pages[$i]["OUT_SCORE"] += $alpha / (59 + $score);
}
}
orderCallback($pages[0], $pages[0], "OUT_SCORE");
}
usort($pages, "orderCallback");
if ($use_proximity) {
for ($i = 0; $i < $result_count; $i++) {
$pages[$i][self::SCORE] = $pages[$i]["OUT_SCORE"];
}
} else {
for ($i = 0; $i < $result_count; $i++) {
$pages[$i][self::PROXIMITY] = 1;
$pages[$i][self::SCORE] = $pages[$i]["OUT_SCORE"];
}
}
$sort_time = changeInMicrotime($sort_start);
}
if ($num_retrieved < $to_retrieve) {
$results['TOTAL_ROWS'] = $num_retrieved;
} else {
$results['TOTAL_ROWS'] = $query_iterator->num_docs;
//this is only an approximation
}
if ($raw == 1 && $save_timestamp_name == "") {
$pages = array_slice($pages, $start_slice);
$pages = array_slice($pages, $limit - $start_slice, $num);
$results['PAGES'] =& $pages;
if ($old_to_retrieve != $to_retrieve) {
$results['HARD_QUERY'] = $old_to_retrieve;
}
return $results;
}
if (QUERY_STATISTICS) {
$this->query_info['QUERY'] .= "{$in2}<b>Lookup Offsets Time</b>: " . changeInMicrotime($lookup_time) . "<br />";
$machine_times = AnalyticsManager::get("MACHINE_TIMES");
if ($machine_times) {
$this->query_info['QUERY'] .= "{$in3}<i>Machine Sub-Times</i>:<br />" . $machine_times . "<br />";
}
$net_times = AnalyticsManager::get("NET_TIMES");
$max_machine_times = AnalyticsManager::get("MAX_MACHINE_TIMES");
if ($net_times && $max_machine_times) {
$this->query_info['QUERY'] .= "{$in3}<i>Network Overhead Sub-Time</i>: " . ($net_times - $max_machine_times) . "<br />";
}
if ($sort_time) {
$this->query_info['QUERY'] .= "{$in3}<i>Merge-Rank Sub-Time</i>: " . $sort_time . "<br />";
}
$summaries_time = microtime();
}
$get_pages = array_slice($pages, $limit, $num);
$to_get_count = count($get_pages);
$groups_with_docs = false;
if (preg_match("/\\bsite:doc\\b/", $original_query)) {
$groups_with_docs = true;
}
$out_pages = array();
$cur_limit = $limit;
while (count($out_pages) < $to_get_count && $get_pages) {
$out_pages = array_merge($out_pages, $this->getSummariesFromOffsets($get_pages, $queue_servers, $raw, $groups_with_docs));
if ($save_timestamp_name != "") {
break;
}
$cur_limit += $num;
$get_pages = array_slice($pages, $cur_limit, $num);
}
$out_pages = array_slice($out_pages, 0, $num);
if (QUERY_STATISTICS) {
$summary_times_string = AnalyticsManager::get("SUMMARY_TIMES");
if ($summary_times_string) {
$round_summary_times = unserialize($summary_times_string);
$summary_delta_time = changeInMicrotime($summaries_time);
$summary_time_info = "{$summary_delta_time}<br /> {$in4}";
$sum_max_time = 0;
foreach ($round_summary_times as $summary_times) {
$i = 0;
$max_time = 0;
foreach ($summary_times as $summary_time) {
$summary_time_info .= "ID_{$i}: " . $summary_time . "{$indent}";
$max_time = $summary_time > $max_time ? $summary_time : $max_time;
$i++;
}
$sum_max_time += $max_time;
}
$net_overhead = $summary_delta_time - $sum_max_time;
$summary_time_info .= "<br />{$in3}<i>Network Overhead Sub-Time</i>: " . $net_overhead;
} else {
$summary_time_info = changeInMicrotime($summaries_time);
}
$this->query_info['QUERY'] .= "{$in2}<b>Get Summaries Time</b>: " . $summary_time_info . "<br />";
}
$results['PAGES'] =& $out_pages;
$results['TIME'] = time();
$lang = guessLocaleFromString($original_query);
$tokenizer = PhraseParser::getTokenizer($lang);
//only use tokenizer if no meta word or disjuncts in query
if (!preg_match('/(\\||\\:)/u', $original_query) && $tokenizer && method_exists($tokenizer, "scoredThesaurusMatches") && method_exists($tokenizer, "tagPartsOfSpeechPhrase") && isset($tokenizer->use_thesaurus)) {
$results = $this->sortByThesaurusScore($results, $original_query, $lang);
}
if (USE_CACHE && $save_timestamp_name == "") {
$CACHE->set($summary_hash, $results);
}
return $results;
}
/**
* In a multiple queue server setting, gets summaries for a set of document
* by their url, or by group of 5-tuples of the form
* (machine, key, index, generation, offset). This makes an execMachines
* call to make a network request to the CrawlController's on each machine
* which in turn calls getCrawlItems (and thence nonNetworkGetCrawlItems)
* on each machine. The results are then sent back to networkGetCrawlItems
* and aggregated.
*
* @param string $lookups things whose summaries we are trying to look up
* @param array $machine_urls an array of urls of yioop queue servers
* @return array of summary data for the matching documents
*/
function networkGetCrawlItems($lookups, $machine_urls)
{
//Set-up network request
$machines = array();
$indexes = array();
$num_machines = count($machine_urls);
foreach ($lookups as $lookup => $lookup_info) {
if (count($lookup_info) == 2 && ($lookup_info[0][0] === 'h' || $lookup_info[0][0] === 'r' || $lookup_info[0][0] === 'g')) {
$machines = $machine_urls;
break;
} else {
foreach ($lookup_info as $lookup_item) {
$out_lookup_info = array();
if (count($lookup_item) == 5) {
list($index, , , , ) = $lookup_item;
$machines[$index] = $machine_urls[$index];
} else {
$machines = $machine_urls;
break;
}
}
}
}
//Make request
$page_set = $this->execMachines("getCrawlItems", $machines, serialize($lookups), $num_machines);
//Aggregate results
$summaries = array();
$elapsed_times = array();
if (is_array($page_set)) {
foreach ($page_set as $elt) {
$description_hash = array();
$result = @unserialize(webdecode($elt[self::PAGE]));
if (!is_array($result)) {
$elapsed_times[] = 0;
continue;
}
$elapsed_times[] = $result["ELAPSED_TIME"];
unset($result["ELAPSED_TIME"]);
$ellipsis = "";
foreach ($result as $lookup => $summary) {
if (isset($summaries[$lookup])) {
if (isset($summary[self::DESCRIPTION])) {
$description = trim($summary[self::DESCRIPTION]);
if (!isset($summaries[$lookup][self::DESCRIPTION])) {
$summaries[$lookup][self::DESCRIPTION] = "";
}
if (!isset($description_hash[$description])) {
$summaries[$lookup][self::DESCRIPTION] = $ellipsis . $description;
$ellipsis = " .. ";
$description_hash[$description] = true;
}
}
foreach ($summary as $attr => $value) {
if ($attr != self::DESCRIPTION && !isset($summaries[$lookup][$attr])) {
$summaries[$lookup][$attr] = $value;
}
}
} else {
$summaries[$lookup] = $summary;
}
}
}
$summary_times_string = AnalyticsManager::get("SUMMARY_TIMES");
if ($summary_times_string) {
$all_elapsed_times = unserialize($summary_times_string);
} else {
$all_elapsed_times = array();
}
$all_elapsed_times[] = $elapsed_times;
AnalyticsManager::set("SUMMARY_TIMES", serialize($all_elapsed_times));
}
return $summaries;
}
/**
* Send the provided view to output, drawing it with the given
* data variable, using the current locale for translation, and
* writing mode
*
* @param string $view the name of the view to draw
* @param array $data an array of values to use in drawing the view
*/
function displayView($view, $data)
{
$data['LOCALE_TAG'] = getLocaleTag();
$data['LOCALE_DIR'] = getLocaleDirection();
$data['BLOCK_PROGRESSION'] = getBlockProgression();
$data['WRITING_MODE'] = getWritingMode();
if (QUERY_STATISTICS) {
$data['QUERY_STATISTICS'] = array();
$machine = isset($_SERVER["HTTP_HOST"]) ? htmlentities($_SERVER["HTTP_HOST"]) : "localhost";
$machine_uri = isset($_SERVER['REQUEST_URI']) ? htmlentities($_SERVER['REQUEST_URI']) : "/";
$protocol = isset($_SERVER["HTTPS"]) ? "https://" : "http://";
if ($machine == '::1') {
//IPv6 :(
$machine = "[::1]/";
//used if the fetching and queue serving on the same machine
}
$data['YIOOP_INSTANCE'] = $protocol . $machine . $machine_uri;
$data['TOTAL_ELAPSED_TIME'] = 0;
foreach ($this->model_instances as $model_name => $model) {
$data['QUERY_STATISTICS'] = array_merge($model->db->query_log, $data['QUERY_STATISTICS']);
$data['TOTAL_ELAPSED_TIME'] += $model->db->total_time;
}
$locale_info = getLocaleQueryStatistics();
$data['QUERY_STATISTICS'] = array_merge($locale_info['QUERY_LOG'], $data['QUERY_STATISTICS']);
$data['TOTAL_ELAPSED_TIME'] += $locale_info['TOTAL_ELAPSED_TIME'];
$mail_total_time = AnalyticsManager::get("MAIL_TOTAL_TIME");
$mail_messages = AnalyticsManager::get("MAIL_MESSAGES");
if ($mail_total_time && $mail_messages) {
$data['QUERY_STATISTICS'] = array_merge($mail_messages, $data['QUERY_STATISTICS']);
$data['TOTAL_ELAPSED_TIME'] += $mail_total_time;
}
}
$data['c'] = isset($_REQUEST['c']) ? $_REQUEST['c'] : NULL;
if (isset($_SESSION['DISPLAY_MESSAGE'])) {
$data['DISPLAY_MESSAGE'] = $_SESSION['DISPLAY_MESSAGE'];
unset($_SESSION['DISPLAY_MESSAGE']);
}
$this->view($view)->render($data);
}
/**
* Hook function used by currentDocsWithWord to return the current block
* of docs if it is not cached
*
* @return mixed doc ids and score if there are docs left, -1 otherwise
*/
function findDocsWithWord()
{
$query = $this->base_query . "&num={$this->results_per_block}&limit={$this->limit}";
$sites = array();
$lookup = array();
$i = 0;
$j = 0;
foreach ($this->queue_servers as $server) {
if ($this->more_flags[$i]) {
$sites[$j][CrawlConstants::URL] = $server . "?" . $query . "&machine={$i}";
$lookup[$j] = $i;
$j++;
}
$i++;
}
$net_times = AnalyticsManager::get("NET_TIMES");
$net_times = $net_times ? $net_times : 0;
$download_time = microtime();
$downloads = array();
if (count($sites) > 0) {
$downloads = FetchUrl::getPages($sites, false, 0, NULL, self::URL, self::PAGE, true);
}
$net_times += changeInMicrotime($download_time);
AnalyticsManager::set("NET_TIMES", $net_times);
$results = array();
$count = count($downloads);
$this->num_docs = 0;
$in4 = " ";
$machine_times = AnalyticsManager::get("MACHINE_TIMES");
$indent = $machine_times ? "<br />{$in4}" : $in4;
$machine_times = $machine_times ? $machine_times : "";
$max_machine_times = AnalyticsManager::get("MAX_MACHINE_TIMES");
$max_machine_times = $max_machine_times ? $max_machine_times : 0;
$max_time = 0;
$num_with_results = $count;
for ($j = 0; $j < $count; $j++) {
$download =& $downloads[$j];
if (isset($download[self::PAGE])) {
$pre_result = @unserialize($download[self::PAGE]);
if (!isset($pre_result["TOTAL_ROWS"]) || $pre_result["TOTAL_ROWS"] < $this->results_per_block) {
$this->more_flags[$lookup[$j]] = false;
$num_with_results--;
}
if (isset($pre_result["TOTAL_ROWS"])) {
$this->num_docs += $pre_result["TOTAL_ROWS"];
}
if (isset($pre_result["PAGES"])) {
foreach ($pre_result["PAGES"] as $page_data) {
if (isset($page_data[self::KEY])) {
$results[$page_data[self::KEY]] = $page_data;
$results[$page_data[self::KEY]][self::MACHINE_ID] = $lookup[$j];
}
}
}
$max_time = max($max_time, $pre_result['ELAPSED_TIME']);
$lookup_link = $this->makeLookupLink($sites, $lookup[$j]);
$machine_times .= $indent . $lookup_link . $pre_result['ELAPSED_TIME'] . " ";
$indent = "";
}
}
if (isset($pre_result["HARD_QUERY"])) {
$this->hard_query = $pre_result["HARD_QUERY"];
}
if ($num_with_results > 0) {
$this->next_results_per_block = ceil(floatval($count * $this->results_per_block) / floatval($num_with_results));
}
$max_machine_times += $max_time;
AnalyticsManager::set("MACHINE_TIMES", $machine_times);
AnalyticsManager::set("MAX_MACHINE_TIMES", $max_machine_times);
if ($results == array()) {
$results = -1;
}
if ($results != -1) {
if ($this->filter != NULL) {
foreach ($results as $keys => $data) {
$host_key = substr($keys, self::HOST_KEY_POS, self::KEY_LEN);
if (in_array($host_key, $this->filter)) {
unset($results[$keys]);
}
}
}
}
$this->count_block = count($results);
$this->pages = $results;
return $results;
}