"Search", "version" => "0.6.2", "author" => "Starbeamrainbowlabs", "description" => "Adds proper search functionality to Pepperminty Wiki using an inverted index to provide a full text search engine. If pages don't show up, then you might have hit a stop word. If not, try requesting the `invindex-rebuild` action to rebuild the inverted index from scratch.", "id" => "feature-search", "code" => function() { global $settings; /** * @api {get} ?action=index&page={pageName} Get an index of words for a given page * @apiName SearchIndex * @apiGroup Search * @apiPermission Anonymous * @apiDescription For debugging purposes. Be warned - the format could change at any time! * * @apiParam {string} page The page to generate a word index page. */ /* * ██ ███ ██ ██████ ███████ ██ ██ * ██ ████ ██ ██ ██ ██ ██ ██ * ██ ██ ██ ██ ██ ██ █████ ███ * ██ ██ ██ ██ ██ ██ ██ ██ ██ * ██ ██ ████ ██████ ███████ ██ ██ */ add_action("index", function() { global $settings, $env; $breakable_chars = "\r\n\t .,\\/!\"£$%^&*[]()+`_~#"; header("content-type: text/plain"); $source = file_get_contents("$env->storage_prefix$env->page.md"); $index = search::index($source); var_dump($env->page); var_dump($source); var_dump($index); }); /** * @api {get} ?action=invindex-rebuild Rebuild the inverted search index from scratch * @apiDescription Causes the inverted search index to be completely rebuilt from scratch. Can take a while for large wikis! * @apiName SearchInvindexRebuild * @apiGroup Search * @apiPermission Admin * * @apiParam {string} secret Optional. Specify the secret from peppermint.json here in order to rebuild the search index without logging in. */ /* * ██ ███ ██ ██ ██ ██ ███ ██ ██████ ███████ ██ ██ * ██ ████ ██ ██ ██ ██ ████ ██ ██ ██ ██ ██ ██ * ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ █████ ███ █████ * ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ * ██ ██ ████ ████ ██ ██ ████ ██████ ███████ ██ ██ * * ██████ ███████ ██████ ██ ██ ██ ██ ██████ * ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ * ██████ █████ ██████ ██ ██ ██ ██ ██ ██ * ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ * ██ ██ ███████ ██████ ██████ ██ ███████ ██████ */ add_action("invindex-rebuild", function() { global $env, $settings; if($env->is_admin || ( !empty($_POST["secret"]) && $_POST["secret"] === $settings->secret ) ) search::rebuild_invindex(); else { http_response_code(401); exit(page_renderer::render_main("Error - Search index regenerator - $settings->sitename", "

Error: You aren't allowed to regenerate the search index. Try logging in as an admin, or setting the secret POST parameter to $settings->sitename's secret - which can be found in $settings->sitename's peppermint.json file.

")); } }); /** * @api {get} ?action=idindex-show Show the id index * @apiDescription Outputs the id index. Useful if you need to verify that it's working as expected. Output is a json object. * @apiName SearchShowIdIndex * @apiGroup Search * @apiPermission Anonymous */ add_action("idindex-show", function() { global $idindex; header("content-type: application/json; charset=UTF-8"); exit(json_encode($idindex, JSON_PRETTY_PRINT)); }); /** * @api {get} ?action=search&query={text}[&format={format}] Search the wiki for a given query string * @apiName Search * @apiGroup Search * @apiPermission Anonymous * * @apiParam {string} query The query string to search for. * @apiParam {string} format Optional. Valid values: html, json. In json mode an object is returned with page names as keys, values as search result information - sorted in ranking order. */ /* * ███████ ███████ █████ ██████ ██████ ██ ██ * ██ ██ ██ ██ ██ ██ ██ ██ ██ * ███████ █████ ███████ ██████ ██ ███████ * ██ ██ ██ ██ ██ ██ ██ ██ ██ * ███████ ███████ ██ ██ ██ ██ ██████ ██ ██ */ add_action("search", function() { global $settings, $env, $pageindex, $paths; // Create the inverted index if it doesn't exist. // todo In the future perhaps a CLI for this would be good? if(!file_exists($paths->searchindex)) search::rebuild_invindex(false); if(!isset($_GET["query"])) exit(page_renderer::render("No Search Terms - Error - $settings->sitename", "

You didn't specify any search terms. Try typing some into the box above.

")); $search_start = microtime(true); $time_start = microtime(true); $invindex = search::load_invindex($paths->searchindex); $env->perfdata->invindex_decode_time = round((microtime(true) - $time_start)*1000, 3); $start = microtime(true); $results = search::query_invindex($_GET["query"], $invindex); $resultCount = count($results); $env->perfdata->invindex_query_time = round((microtime(true) - $time_start)*1000, 3); header("x-invindex-decode-time: {$env->perfdata->invindex_decode_time}ms"); header("x-invindex-query-time: {$env->perfdata->invindex_query_time}ms"); foreach($results as &$result) { $result["context"] = search::extract_context( $_GET["query"], file_get_contents($env->storage_prefix . $result["pagename"] . ".md") ); } if(!empty($_GET["format"]) && $_GET["format"] == "json") { header("content-type: application/json"); $json_results = new stdClass(); foreach($results as $result) $json_results->{$result["pagename"]} = $result; exit(json_encode($json_results)); } $env->perfdata->search_time = round((microtime(true) - $search_start)*1000, 3); $title = $_GET["query"] . " - Search results - $settings->sitename"; $content = "
\n"; $content .= "

Search Results

"; /// Search Box /// $content .= "
\n"; $content .= " \n"; $content .= " \n"; $content .= "
"; $content .= "

Found $resultCount " . ($resultCount === 1 ? "result" : "results") . " in " . $env->perfdata->search_time . "ms. "; $query = $_GET["query"]; if(isset($pageindex->$query)) { $content .= "There's a page on $settings->sitename called $query."; } else { $content .= "There isn't a page called $query on $settings->sitename, but you "; if((!$settings->anonedits && !$env->is_logged_in) || !$settings->editing) { $content .= "do not have permission to create it."; if(!$env->is_logged_in) { $content .= " You could try logging in."; } } else { $content .= "can create it."; } } $content .= "

"; if(module_exists("page-list")) { $nterms = search::tokenize($query); $nterms_regex = implode("|", array_map(function($nterm) { return preg_quote(strtolower(trim($nterm))); }, $nterms)); $all_tags = get_all_tags(); $matching_tags = []; foreach($all_tags as $tag) { if(preg_match("/$nterms_regex/i", trim($tag)) > 0) $matching_tags[] = $tag; } if(count($matching_tags) > 0) { $content .= "

Matching tags: "; foreach($matching_tags as $tag) { $content .= "\t" . htmlentities($tag) . " \n"; } $content .= "

"; } } $i = 0; // todo use $_GET["offset"] and $_GET["result-count"] or something foreach($results as $result) { $link = "?page=" . rawurlencode($result["pagename"]); $pagesource = file_get_contents($env->storage_prefix . $result["pagename"] . ".md"); //echo("Extracting context for result " . $result["pagename"] . ".\n"); $context = $result["context"]; if(mb_strlen($context) === 0) $context = mb_substr($pagesource, 0, $settings->search_characters_context * 2); //echo("'Generated search context for " . $result["pagename"] . ": $context'\n"); $context = search::highlight_context( $_GET["query"], preg_replace('/page.md", null, null, null, $settings->search_characters_context * 2)); if($pageindex->{$env->page}->size > $settings->search_characters_context * 2) $context .= "..."; }*/ $tag_list = ""; foreach($pageindex->{$result["pagename"]}->tags ?? [] as $tag) $tag_list .= "$tag"; $tag_list .= "\n"; // Make redirect pages italics if(!empty($pageindex->{$result["pagename"]}->redirect)) $result["pagename"] = "{$result["pagename"]}"; // We add 1 to $i here to convert it from an index to a result // number as people expect it to start from 1 $content .= "
\n"; $content .= "

" . $result["pagename"] . " $tag_list

\n"; $content .= "

$context

\n"; $content .= "
\n"; $i++; } $content .= "
\n"; header("content-type: text/html; charset=UTF-8"); exit(page_renderer::render($title, $content)); //header("content-type: text/plain"); //var_dump($results); }); /* * ██████ ██ ██ ███████ ██████ ██ ██ * ██ ██ ██ ██ ██ ██ ██ ██ ██ * ██ ██ ██ ██ █████ ██████ ████ █████ * ██ ▄▄ ██ ██ ██ ██ ██ ██ ██ * ██████ ██████ ███████ ██ ██ ██ * ▀▀ * ███████ ███████ █████ ██████ ██████ ██ ██ ██ ███ ██ ██████ ███████ ██ ██ * ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ████ ██ ██ ██ ██ ██ ██ * ███████ █████ ███████ ██████ ██ ███████ ██ ██ ██ ██ ██ ██ █████ ███ * ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ * ███████ ███████ ██ ██ ██ ██ ██████ ██ ██ ██ ██ ████ ██████ ███████ ██ ██ */ /** * @api {get} ?action=query-searchindex&query={text} Inspect the internals of the search results for a query * @apiName Search * @apiGroup Search * @apiPermission Anonymous * * @apiParam {string} query The query string to search for. */ add_action("query-searchindex", function() { global $env, $paths; if(empty($_GET["query"])) { http_response_code(400); header("content-type: text/plain"); exit("Error: No query specified. Specify it with the 'query' GET parameter."); } $env->perfdata->searchindex_decode_start = microtime(true); $searchIndex = search::load_invindex($paths->searchindex); $env->perfdata->searchindex_decode_time = (microtime(true) - $env->perfdata->searchindex_decode_start) * 1000; $env->perfdata->searchindex_query_start = microtime(true); $searchResults = search::query_invindex($_GET["query"], $searchIndex); $env->perfdata->searchindex_query_time = (microtime(true) - $env->perfdata->searchindex_query_start) * 1000; header("content-type: application/json"); $result = new stdClass(); $result->time_format = "ms"; $result->decode_time = $env->perfdata->searchindex_decode_time; $result->query_time = $env->perfdata->searchindex_query_time; $result->total_time = $result->decode_time + $result->query_time; $result->search_results = $searchResults; exit(json_encode($result, JSON_PRETTY_PRINT)); }); /* * ██████ ██████ ███████ ███ ██ ███████ ███████ █████ ██████ ██████ ██ ██ * ██ ██ ██ ██ ██ ████ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ * ██ ██ ██████ █████ ██ ██ ██ ███████ █████ ███████ ██████ ██ ███████ * ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ * ██████ ██ ███████ ██ ████ ███████ ███████ ██ ██ ██ ██ ██████ ██ ██ */ /** * @api {get} ?action=opensearch-description Get the opensearch description file * @apiName OpenSearchDescription * @apiGroup Search * @apiPermission Anonymous */ add_action("opensearch-description", function () { global $settings; $siteRoot = full_url() . "/index.php"; if(!isset($_GET["debug"])) header("content-type: application/opensearchdescription+xml"); else header("content-type: text/plain"); exit('' . // hack The build system strips it otherwise O.o I should really fix that. "\n Search $settings->sitename Search $settings->sitename, which is powered by Pepperminty Wiki. $settings->sitename Wiki $settings->favicon Search content available under the license linked to at the bottom of the search results page. Starbeamrainbowlabs (https://github.com/sbrl/Pepperminty-Wiki/graphs/contributors) UTF-8 UTF-8 "); }); /** * @api {get} ?action=suggest-pages[&type={type}] Get page name suggestions for a query * @apiName OpenSearchDescription * @apiGroup Search * @apiPermission Anonymous * * @apiParam {string} text The search query string to get search suggestions for. * @apiParam {string} type The type of result to return. Default value: json. Available values: json, opensearch */ add_action("suggest-pages", function() { global $settings, $pageindex; if($settings->dynamic_page_suggestion_count === 0) { header("content-type: application/json"); header("content-length: 3"); exit("[]\n"); } if(empty($_GET["query"])) { http_response_code(400); header("content-type: text/plain"); exit("Error: You didn't specify the 'query' GET parameter."); } $type = $_GET["type"] ?? "json"; if(!in_array($type, ["json", "opensearch"])) { http_response_code(406); exit("Error: The type '$type' is not one of the supported output types. Available values: json, opensearch. Default: json"); } $literator = Transliterator::createFromRules(':: Any-Latin; :: Latin-ASCII; :: NFD; :: [:Nonspacing Mark:] Remove; :: Lower(); :: NFC;', Transliterator::FORWARD); $query = $literator->transliterate($_GET["query"]); // Rank each page name $results = []; foreach($pageindex as $pageName => $entry) { $results[] = [ "pagename" => $pageName, // Costs: Insert: 1, Replace: 8, Delete: 6 "distance" => levenshtein($query, $literator->transliterate($pageName), 1, 8, 6) ]; } // Sort the page names by distance from the original query usort($results, function($a, $b) { if($a["distance"] == $b["distance"]) return strcmp($a["pagename"], $b["pagename"]); return $a["distance"] < $b["distance"] ? -1 : 1; }); // Send the results to the user $suggestions = array_slice($results, 0, $settings->dynamic_page_suggestion_count); switch($type) { case "json": header("content-type: application/json"); exit(json_encode($suggestions)); case "opensearch": $opensearch_output = [ $_GET["query"], array_map(function($suggestion) { return $suggestion["pagename"]; }, $suggestions) ]; header("content-type: application/x-suggestions+json"); exit(json_encode($opensearch_output)); } }); if($settings->dynamic_page_suggestion_count > 0) { page_renderer::AddJSSnippet('/// Dynamic page suggestion system // Micro snippet 8 - Promisified GET (fetched 20th Nov 2016) function get(u){return new Promise(function(r,t,a){a=new XMLHttpRequest();a.onload=function(b,c){b=a.status;c=a.response;if(b>199&&b<300){r(c)}else{t(c)}};a.open("GET",u,true);a.send(null)})} window.addEventListener("load", function(event) { var searchBox = document.querySelector("input[type=search]"); searchBox.dataset.lastValue = ""; searchBox.addEventListener("keyup", function(event) { // Make sure that we don\'t keep sending requests to the server if nothing has changed if(searchBox.dataset.lastValue == event.target.value) return; searchBox.dataset.lastValue = event.target.value; // Fetch the suggestions from the server get("?action=suggest-pages&query=" + encodeURIComponent(event.target.value)).then(function(response) { var suggestions = JSON.parse(response), dataList = document.getElementById("allpages"); // If the server sent no suggestions, then we shouldn\'t replace the contents of the datalist if(suggestions.length == 0) return; console.info(`Fetched suggestions for ${event.target.value}:`, suggestions.map(s => s.pagename)); // Remove all the existing suggestions while(dataList.firstChild) { dataList.removeChild(dataList.firstChild); } // Add the new suggestions to the datalist var optionsFrag = document.createDocumentFragment(); suggestions.forEach(function(suggestion) { var suggestionElement = document.createElement("option"); suggestionElement.value = suggestion.pagename; suggestionElement.dataset.distance = suggestion.distance; optionsFrag.appendChild(suggestionElement); }); dataList.appendChild(optionsFrag); }); }); }); '); } } ]); /** * Holds a collection to methods to manipulate various types of search index. */ class search { /** * Words that we should exclude from the inverted index */ public static $stop_words = [ "a", "about", "above", "above", "across", "after", "afterwards", "again", "against", "all", "almost", "alone", "along", "already", "also", "although", "always", "am", "among", "amongst", "amoungst", "amount", "an", "and", "another", "any", "anyhow", "anyone", "anything", "anyway", "anywhere", "are", "around", "as", "at", "back", "be", "became", "because", "become", "becomes", "becoming", "been", "before", "beforehand", "behind", "being", "below", "beside", "besides", "between", "beyond", "bill", "both", "bottom", "but", "by", "call", "can", "cannot", "cant", "co", "con", "could", "couldnt", "cry", "de", "describe", "detail", "do", "done", "down", "due", "during", "each", "eg", "eight", "either", "eleven", "else", "elsewhere", "empty", "enough", "etc", "even", "ever", "every", "everyone", "everything", "everywhere", "except", "few", "fill", "find", "fire", "first", "five", "for", "former", "formerly", "found", "four", "from", "front", "full", "further", "get", "give", "go", "had", "has", "hasnt", "have", "he", "hence", "her", "here", "hereafter", "hereby", "herein", "hereupon", "hers", "herself", "him", "himself", "his", "how", "however", "ie", "if", "in", "inc", "indeed", "interest", "into", "is", "it", "its", "itself", "keep", "last", "latter", "latterly", "least", "less", "ltd", "made", "many", "may", "me", "meanwhile", "might", "mine", "more", "moreover", "most", "mostly", "move", "much", "must", "my", "myself", "name", "namely", "neither", "never", "nevertheless", "next", "nine", "no", "none", "nor", "not", "nothing", "now", "nowhere", "of", "off", "often", "on", "once", "one", "only", "onto", "or", "other", "others", "otherwise", "our", "ours", "ourselves", "out", "over", "own", "part", "per", "perhaps", "please", "put", "rather", "re", "same", "see", "seem", "seemed", "seeming", "seems", "serious", "several", "she", "should", "show", "side", "since", "sincere", "six", "sixty", "so", "some", "somehow", "someone", "something", "sometime", "sometimes", "somewhere", "still", "such", "system", "take", "ten", "than", "that", "the", "their", "them", "themselves", "then", "thence", "there", "thereafter", "thereby", "therefore", "therein", "thereupon", "these", "they", "thickv", "thin", "third", "this", "those", "though", "three", "through", "throughout", "thru", "thus", "to", "together", "too", "top", "toward", "towards", "twelve", "twenty", "two", "un", "under", "until", "up", "upon", "us", "very", "via", "was", "we", "well", "were", "what", "whatever", "when", "whence", "whenever", "where", "whereafter", "whereas", "whereby", "wherein", "whereupon", "wherever", "whether", "which", "while", "whither", "who", "whoever", "whole", "whom", "whose", "why", "will", "with", "within", "without", "would", "yet", "you", "your", "yours", "yourself", "yourselves" ]; /** * Converts a source string into an index of search terms that can be * merged into an inverted index. * @param string $source The source string to index. * @return array An index represents the specified string. */ public static function index($source) { // We don't need to normalise or transliterate here because self::tokenize() does this for us $source = html_entity_decode($source, ENT_QUOTES); $source_length = mb_strlen($source); $index = []; $terms = self::tokenize($source); $i = 0; foreach($terms as $term) { $nterm = $term; // Skip over stop words (see https://en.wikipedia.org/wiki/Stop_words) if(in_array($nterm, self::$stop_words)) continue; if(!isset($index[$nterm])) { $index[$nterm] = [ "freq" => 0, "offsets" => [] ]; } $index[$nterm]["freq"]++; $index[$nterm]["offsets"][] = $i; $i++; } return $index; } /** * Converts a source string into a series of raw tokens. * @param string $source The source string to process. * @return array An array of raw tokens extracted from the specified source string. */ public static function tokenize($source) { /** Normalises input characters for searching & indexing */ static $literator; $literator = Transliterator::createFromRules(':: Any-Latin; :: Latin-ASCII; :: NFD; :: [:Nonspacing Mark:] Remove; :: Lower(); :: NFC;', Transliterator::FORWARD); // We don't need to normalise here because the transliterator handles // this for us. Also, we can't move the literator to a static variable // because PHP doesn't like it very much $source = $literator->transliterate($source); $source = preg_replace('/[\[\]\|\{\}\/]/u', " ", $source); return preg_split("/((^\p{P}+)|(\p{P}*\s+\p{P}*)|(\p{P}+$))|\|/u", $source, -1, PREG_SPLIT_NO_EMPTY); } /** * Removes (most) markdown markup from the specified string. * Stripped strings are not suitable for indexing! * @param string $source The source string to process. * @return string The stripped string. */ public static function strip_markup($source) { return preg_replace('/([\"*_\[\]]| - |`)/u', "", $source); } /** * Rebuilds the master inverted index and clears the page id index. * @param boolean $output Whether to send progress information to the user's browser. */ public static function rebuild_invindex($output = true) { global $pageindex, $env, $paths, $settings; /** Normalises input characters for searching & indexing */ static $literator; $literator = Transliterator::createFromRules(':: Any-Latin; :: Latin-ASCII; :: NFD; :: [:Nonspacing Mark:] Remove; :: Lower(); :: NFC;', Transliterator::FORWARD); if($output) { header("content-type: text/event-stream"); ob_end_flush(); } // Clear the id index out ids::clear(); // Reindex each page in turn $invindex = []; $i = 0; $max = count(get_object_vars($pageindex)); $missing_files = 0; foreach($pageindex as $pagename => $pagedetails) { $page_filename = $env->storage_prefix . $pagedetails->filename; if(!file_exists($page_filename)) { echo("data: [" . ($i + 1) . " / $max] Error: Can't find $page_filename\n"); flush(); $i++; $missing_files++; continue; } // We do not transliterate or normalise here because the indexer will take care of this for us $index = self::index(file_get_contents($page_filename)); $pageid = ids::getid($pagename); self::merge_into_invindex($invindex, $pageid, $index); if($output) { echo("data: [" . ($i + 1) . " / $max] Added $pagename (id #$pageid) to the new search index.\n\n"); flush(); } $i++; } if($output) { echo("data: Search index rebuilding complete.\n\n"); echo("data: Couldn't find $missing_files pages on disk. If $settings->sitename couldn't find some pages on disk, then you might need to manually correct $settings->sitename's page index (stored in pageindex.json).\n\n"); echo("data: Done! Saving new search index to '$paths->searchindex'.\n\n"); } self::save_invindex($paths->searchindex, $invindex); } /** * Sorts an index alphabetically. Will also sort an inverted index. * This allows us to do a binary search instead of a regular * sequential search. * @param array $index The index to sort. */ public static function sort_index(&$index) { ksort($index, SORT_NATURAL); } /** * Compares two *regular* indexes to find the differences between them. * @param array $oldindex The old index. * @param array $newindex The new index. * @param array $changed An array to be filled with the nterms of all the changed entries. * @param array $removed An array to be filled with the nterms of all the removed entries. */ public static function compare_indexes($oldindex, $newindex, &$changed, &$removed) { foreach($oldindex as $nterm => $entry) { if(!isset($newindex[$nterm])) $removed[] = $nterm; } foreach($newindex as $nterm => $entry) { if(!isset($oldindex[$nterm]) or // If this world is new $newindex[$nterm] !== $oldindex[$nterm]) // If this word has changed $changed[$nterm] = $newindex[$nterm]; } } /** * Reads in and parses an inverted index. * @param string $invindex_filename The path tp the inverted index to parse. * @todo Remove this function and make everything streamable */ public static function load_invindex($invindex_filename) { $invindex = json_decode(file_get_contents($invindex_filename), true); return $invindex; } /** * Reads in and parses an inverted index, measuring the time it takes to do so. * @param string $invindex_filename The path to the file inverted index to parse. */ public static function measure_invindex_load_time($invindex_filename) { global $env; $searchindex_decode_start = microtime(true); search::load_invindex($invindex_filename); $env->perfdata->searchindex_decode_time = round((microtime(true) - $searchindex_decode_start)*1000, 3); } /** * Merge an index into an inverted index. * @param array $invindex The inverted index to merge into. * @param int $pageid The id of the page to assign to the index that's being merged. * @param array $index The regular index to merge. * @param array $removals An array of index entries to remove from the inverted index. Useful for applying changes to an inverted index instead of deleting and remerging an entire page's index. */ public static function merge_into_invindex(&$invindex, $pageid, &$index, &$removals = []) { // Remove all the subentries that were removed since last time foreach($removals as $nterm) { unset($invindex[$nterm][$pageid]); } // Merge all the new / changed index entries into the inverted index foreach($index as $nterm => $newentry) { // If the nterm isn't in the inverted index, then create a space for it if(!isset($invindex[$nterm])) $invindex[$nterm] = []; $invindex[$nterm][$pageid] = $newentry; // Sort the page entries for this word by frequency uasort($invindex[$nterm], function($a, $b) { if($a["freq"] == $b["freq"]) return 0; return ($a["freq"] < $b["freq"]) ? +1 : -1; }); } // Sort the inverted index by rank uasort($invindex, function($a, $b) { $ac = count($a); $bc = count($b); if($ac == $bc) return 0; return ($ac < $bc) ? +1 : -1; }); } /** * Deletes the given pageid from the given pageindex. * @param inverted_index &$invindex The inverted index. * @param number $pageid The pageid to remove. */ public static function delete_entry(&$invindex, $pageid) { $str_pageid = (string)$pageid; foreach($invindex as $nterm => &$entry) { if(isset($entry[$pageid])) unset($entry[$pageid]); if(isset($entry[$str_pageid])) unset($entry[$str_pageid]); if(count($entry) === 0) unset($invindex[$nterm]); } } /** * Saves the given inverted index back to disk. * @param string $filename The path to the file to save the inverted index to. * @param array $invindex The inverted index to save. */ public static function save_invindex($filename, &$invindex) { file_put_contents($filename, json_encode($invindex)); } /** * Searches the given inverted index for the specified search terms. * @param string $query The search query. * @param array $invindex The inverted index to search. * @return array An array of matching pages. */ public static function query_invindex($query, &$invindex) { global $settings, $pageindex; /** Normalises input characters for searching & indexing */ static $literator; $literator = Transliterator::createFromRules(':: Any-Latin; :: Latin-ASCII; :: NFD; :: [:Nonspacing Mark:] Remove; :: Lower(); :: NFC;', Transliterator::FORWARD); $query_terms = self::tokenize($query); $matching_pages = []; // Loop over each term in the query and find the matching page entries $count = count($query_terms); for($i = 0; $i < $count; $i++) { $qterm = $query_terms[$i]; // Stop words aren't worth the bother - make sure we don't search // the title or the tags for them if(in_array($qterm, self::$stop_words)) continue; // Only search the inverted index if it actually exists there if(isset($invindex[$qterm])) { // Loop over each page in the inverted index entry reset($invindex); // Reset array/object pointer foreach($invindex[$qterm] as $pageid => $page_entry) { // Create an entry in the matching pages array if it doesn't exist if(!isset($matching_pages[$pageid])) $matching_pages[$pageid] = [ "nterms" => [] ]; $matching_pages[$pageid]["nterms"][$qterm] = $page_entry; } } // Loop over the pageindex and search the titles / tags reset($pageindex); // Reset array/object pointer foreach ($pageindex as $pagename => $pagedata) { // Get the current page's id $pageid = ids::getid($pagename); // Consider matches in the page title $title_matches = mb_stripos_all($literator->transliterate($pagename), $qterm); $title_matches_count = $title_matches !== false ? count($title_matches) : 0; if($title_matches_count > 0) { // We found the qterm in the title if(!isset($matching_pages[$pageid])) $matching_pages[$pageid] = [ "nterms" => [] ]; // Set up a counter for page title matches if it doesn't exist already if(!isset($matching_pages[$pageid]["title-matches"])) $matching_pages[$pageid]["title-matches"] = 0; $matching_pages[$pageid]["title-matches"] += $title_matches_count * strlen($qterm); } // Consider matches in the page's tags $tag_matches = isset($pagedata->tags) ? mb_stripos_all($literator->transliterate(implode(" ", $pagedata->tags)), $qterm) : false; $tag_matches_count = $tag_matches !== false ? count($tag_matches) : 0; if($tag_matches_count > 0) // And we found the qterm in the tags { if(!isset($matching_pages[$pageid])) $matching_pages[$pageid] = [ "nterms" => [] ]; // Set up a counter for tag match if there isn't one already if(!isset($matching_pages[$pageid]["tag-matches"])) $matching_pages[$pageid]["tag-matches"] = 0; $matching_pages[$pageid]["tag-matches"] += $tag_matches_count * strlen($qterm); } } } foreach($matching_pages as $pageid => &$pagedata) { $pagedata["pagename"] = ids::getpagename($pageid); $pagedata["rank"] = 0; $pageOffsets = []; // Loop over each search term found on this page foreach($pagedata["nterms"] as $pterm => $entry) { // Add the number of occurrences of this search term to the ranking // Multiply it by the length of the word $pagedata["rank"] += $entry["freq"] * strlen($pterm); // Add the offsets to a listof all offsets on this page foreach($entry["offsets"] as $offset) $pageOffsets[] = $offset; } /* // Sort the list of offsets $pageOffsets = array_unique($pageOffsets); sort($pageOffsets); var_dump($pageOffsets); // Calcualate the clump distances via a variable moving window size $pageOffsetsCount = count($pageOffsets); $clumpDistanceWindow = min($count, $pageOffsetsCount); // a.k.a. count($query_terms) - see above $clumpDistances = []; for($i = 0; $i < $pageOffsetsCount - $clumpDistanceWindow; $i++) $clumpDistances[] = $pageOffsets[$i] - $pageOffsets[$i + $clumpDistanceWindow]; // Sort the new list of clump distances sort($clumpDistances); // Calcualate a measure of how clumped the offsets are $tightClumpLimit = floor((count($clumpDistances) - 1) / 0.25); $tightClumpsMeasure = $clumpDistances[$tightClumpLimit] - $clumpDistances[0]; $clumpsRange = $clumpDistances[count($clumpDistances) - 1] - $clumpDistances[0]; $clumpiness = $tightClumpsMeasure / $clumpsRange; echo("{$pagedata["pagename"]} - $clumpiness"); */ // Consider matches in the title / tags if(isset($pagedata["title-matches"])) $pagedata["rank"] += $pagedata["title-matches"] * $settings->search_title_matches_weighting; if(isset($pagedata["tag-matches"])) $pagedata["rank"] += $pagedata["tag-matches"] * $settings->search_tags_matches_weighting; // todo remove items if the rank is below a threshold } // todo sort by rank here uasort($matching_pages, function($a, $b) { if($a["rank"] == $b["rank"]) return 0; return ($a["rank"] < $b["rank"]) ? +1 : -1; }); return $matching_pages; } /** * Extracts a context string (in HTML) given a search query that could be displayed * in a list of search results. * @param string $query The search queary to generate the context for. * @param string $source The page source to extract the context from. * @return string The generated context string. */ public static function extract_context($query, $source) { global $settings; $nterms = self::tokenize($query); $matches = []; // Loop over each nterm and find it in the source foreach($nterms as $nterm) { if(in_array($nterm, static::$stop_words)) continue; $all_offsets = mb_stripos_all($source, $nterm); // Skip over adding matches if there aren't any if($all_offsets === false) continue; foreach($all_offsets as $offset) $matches[] = [ $nterm, $offset ]; } // Sort the matches by offset usort($matches, function($a, $b) { if($a[1] == $b[1]) return 0; return ($a[1] > $b[1]) ? +1 : -1; }); $sourceLength = mb_strlen($source); $contexts = []; $matches_count = count($matches); $total_context_length = 0; for($i = 0; $i < $matches_count; $i++) { $next_context = [ "from" => max(0, $matches[$i][1] - $settings->search_characters_context), "to" => min($sourceLength, $matches[$i][1] + count($matches[0]) + $settings->search_characters_context) ]; if(end($contexts) !== false && end($contexts)["to"] > $next_context["from"]) { // This next context overlaps with the previous one // Extend the last one instead of adding a new one // The array pointer is pointing at the last element now because we called end() above // Update the total context length counter appropriately $total_context_length += $next_context["to"] - $contexts[key($contexts)]["to"]; $contexts[key($contexts)]["to"] = $next_context["to"]; } else { // No overlap here! Business as usual. $contexts[] = $next_context; // Update the total context length counter as normal $total_context_length += $next_context["to"] - $next_context["from"]; } end($contexts); $last_context = &$contexts[key($contexts)]; if($total_context_length > $settings->search_characters_context_total) { // We've reached the limit on the number of characters this context should contain. Trim off the context to fit and break out $last_context["to"] -= $total_context_length - $settings->search_characters_context_total; break; } } $contexts_text = []; foreach($contexts as $context) { $contexts_text[] = substr($source, $context["from"], $context["to"] - $context["from"]); } return implode(" ... ", $contexts_text); } /** * Highlights the keywords of a context string. * @param string $query The query to use when highlighting. * @param string $context The context string to highlight. * @return string The highlighted (HTML) string. */ public static function highlight_context($query, $context) { $qterms = self::tokenize($query); foreach($qterms as $qterm) { if(in_array($qterm, static::$stop_words)) continue; // From http://stackoverflow.com/a/2483859/1460422 $context = preg_replace("/" . preg_replace('/\\//u', "\/", preg_quote($qterm)) . "/iu", "$0", $context); } return $context; } } ?>