<table id='uts-mapping-table'>
<tr>
<th>
Uthmani
</th>
<th>
Simple
</th>
</tr>
<?php
$uthmaniCounter = 0;
$simpleCounter = 0;
$qaOntologyConceptsIterator = getAPCIterator("AR\\/OTHERS\\/UTHMANI_TO_SIMPLE_WORD_MAP\\/.*");
foreach ($qaOntologyConceptsIterator as $conceptsCursor) {
$mapTermKey = getEntryKeyFromAPCKey($conceptsCursor['key']);
$mapTermVal = $conceptsCursor['value'];
if (isSimpleQuranWord($mapTermKey)) {
$simpleCounter++;
//echoN("##".$mapTermKey);
continue;
}
$uthmaniCounter++;
?>
<tr>
<td><?php
echo $mapTermKey;
?>
</td>
<td><?php
?>
<table id='pos-options-table'>
<tr>
<td>
Select PoS Tag
</td>
<td>
<select id='pos-selection'>
<option value="" selected="true"> </option>
<?php
$qacPoSTagsIterator = getAPCIterator("AR\\/MODEL_QAC\\/QAC_POS\\/.*");
foreach ($qacPoSTagsIterator as $qacPoSTagsIteratorCursor) {
$posArr = $qacPoSTagsIteratorCursor['value'];
$key = $qacPoSTagsIteratorCursor['key'];
$posTypeName = getEntryKeyFromAPCKey($key);
?>
<option value="<?php
echo $posTypeName;
?>
"><b><?php
echo $posTypeName;
?>
</b></option>
<?php
}
?>
</select>
# ====================================================================
require_once "../global.settings.php";
require_once "../libs/core.lib.php";
loadModels("core,qac", "AR");
$LEMMA_TO_SIMPLE_WORD_MAP = loadLemmaToSimpleMappingTable();
printHTMLPageHeader();
//$qacMasterTableEntryArr2 = getModelEntryFromMemory("AR","MODEL_QAC","QAC_POS",$qacLocation);
$qacPoSTagsIterator = getAPCIterator("AR\\/MODEL_QAC\\/QAC_POS\\/.*");
$QURAN_TEXT = getModelEntryFromMemory("AR", "MODEL_CORE", "QURAN_TEXT", "");
$TOTALS = getModelEntryFromMemory("AR", "MODEL_CORE", "TOTALS", "");
$PAUSEMARKS = $TOTALS['PAUSEMARKS'];
preprint_r($PAUSEMARKS);
foreach ($qacPoSTagsIterator as $qacPoSTagsIteratorCursor) {
$POS_ARR = $qacPoSTagsIteratorCursor['value'];
$key = $qacPoSTagsIteratorCursor['key'];
$POS = getEntryKeyFromAPCKey($key);
if ($POS == "N" || $POS == "PN" || $POS == "ADJ") {
continue;
}
//echoN("|$POS|");
foreach ($POS_ARR as $location => $segmentId) {
$qacMasterTableEntry = getModelEntryFromMemory("AR", "MODEL_QAC", "QAC_MASTERTABLE", $location);
// get Word, Lema and root
$segmentWord = $qacMasterTableEntry[$segmentId - 1]['FORM_AR'];
$segmentWordLema = $qacMasterTableEntry[$segmentId - 1]['FEATURES']['LEM'];
$segmentWordRoot = $qacMasterTableEntry[$segmentId - 1]['FEATURES']['ROOT'];
$verseLocation = substr($location, 0, strlen($location) - 2);
//$segmentWord = removeTashkeel($segmentWord);
if ($POS == "DET") {
// second segment PoS
$segmentPoStag = $qacMasterTableEntry[$segmentId]['TAG'];
#
# You can use Quran Analysis code, framework or corpora in your website
# or application (commercial/non-commercial) provided that you link
# back to www.qurananalysis.com and sufficient credits are given.
#
# ====================================================================
require_once "../global.settings.php";
$lang = "AR";
loadModels("", $lang);
$QURAN_TEXT = getModelEntryFromMemory($lang, "MODEL_CORE", "QURAN_TEXT", "");
preprint_r($QURAN_TEXT[1][1]);
$location = "1:1:1";
$qacMasterTableEntry = getModelEntryFromMemory("AR", "MODEL_QAC", "QAC_MASTERTABLE", $location);
preprint_r($qacMasterTableEntry);
$qaOntologyConceptsIterator = getAPCIterator("ALL\\/MODEL_QA_ONTOLOGY\\/CONCEPTS\\/.*");
foreach ($qaOntologyConceptsIterator as $conceptsCursor) {
$conceptNameID = getEntryKeyFromAPCKey($conceptsCursor['key']);
$conceptArr = $conceptsCursor['value'];
$conceptLabelAR = $conceptArr['label_ar'];
$conceptLabelEN = $conceptArr['label_en'];
$conceptFrequency = $conceptArr['frequency'];
$conceptWeight = $conceptArr['weight'];
preprint_r($conceptArr);
break;
//only one concept
}
// print all words in wordnet
preprint_r(array_keys($MODEL_WORDNET['INDEX']));
// get all information about "Egypt" from wordnet
$wordNetEntry = getWordnetEntryByWordString("egypt");
preprint_r($wordNetEntry);
function isWordPartOfAVerbInVerbIndex($word, $lang)
{
$verbIndexIterator = getAPCIterator("ALL\\/MODEL_QA_ONTOLOGY\\/VERB_INDEX\\/.*");
foreach ($verbIndexIterator as $verbIndexCursor) {
$verbWord = getEntryKeyFromAPCKey($verbIndexCursor['key']);
$verbArr = $verbIndexCursor['value'];
if ($lang == "EN") {
$verbWord = strtolower($verbWord);
}
if (mb_strpos($verbWord, $word) !== false) {
//echoN("|$verbWord| |$word|".( mb_strpos($verbWord, $word)!==false));
return $verbArr;
}
}
return false;
}
function ontologyToD3TreemapHierarchical($MODEL_QA_ONTOLOGY, $minFreq = 0, $lang)
{
global $lang;
$alreadyInLevel1 = array();
$treeRootObj = array();
$treeRootObj["name"] = "قرآن";
if ($lang == "EN") {
$treeRootObj["name"] = "Quran";
}
$treeRootObj["children"] = array();
/** SHOULD BE ZERO BASED FOR D3 TO WORK - o.target.weight = NULL**/
$nodeSerialNumber = 0;
$qaOntologyConceptsIterator = getAPCIterator("ALL\\/MODEL_QA_ONTOLOGY\\/CONCEPTS\\/.*");
foreach ($qaOntologyConceptsIterator as $conceptsCursor) {
$conceptNameID = getEntryKeyFromAPCKey($conceptsCursor['key']);
$conceptArr = $conceptsCursor['value'];
$conceptLabelAR = $conceptArr['label_ar'];
$conceptLabelEN = $conceptArr['label_en'];
$conceptFrequency = $conceptArr['frequency'];
$conceptWeight = $conceptArr['weight'];
$type = $conceptArr['type'];
if ($conceptFrequency < $minFreq) {
continue;
}
//ONLY CLASSES (CLUSTERS)
if ($type == "class") {
$alreadyInLevel1[$conceptNameID] = 1;
//echoN($conceptLabelAR);
//echoN($conceptNameID);
if ($lang == "EN") {
$conceptNameClean = convertConceptIDtoGraphLabel($conceptLabelEN);
} else {
$conceptNameClean = convertConceptIDtoGraphLabel($conceptLabelAR);
}
/*= array("id"=>$nodeSerialNumber++,"word"=>$conceptLabelAR,
"size"=>$conceptWeight,"x"=>rand($startLocationXMin,$startLocationXMax),
"y"=>rand($startLocationYMin,$startLocationYMax));*/
$treeRootObj["children"][] = array("name" => $conceptNameClean, "size" => $conceptFrequency, "children" => getTreeNodeChildren($MODEL_QA_ONTOLOGY, $conceptNameID, $minFreq, $lang, 1, $alreadyInLevel1));
}
}
/*
foreach($MODEL_QA_ONTOLOGY['RELATIONS'] as $index => $relArr)
{
$subject = $relArr['subject'];
$verbAR = $relArr['verb'];
$verbEN = $relArr['VERB_TRANSLATION_EN'];
$verbUthmani = $relArr['verb_uthmani'];
$relFreq = $relArr['frequency'];
$object = $relArr['object'];
//$treeRootObj[$subject]["children"][]["name"]=$object;
$objectConceptArr = $MODEL_QA_ONTOLOGY['CONCEPTS'][$object];
$index = search2DArrayForValue($currentArr,$subject);
$isObjectIncludedBefore = search2DArrayForValue($currentArr[$index]["children"],$object);
if ( $isObjectIncludedBefore===false)
{
//$currentArr[$index]["children"][] = array("name"=>$object,"size"=>$objectConceptArr['frequency'],"children"=>array());
}
}*/
//echoN(count($treeRootObj["children"]));
return $treeRootObj;
}
function loadModels($modelsToBeLoaded, $lang)
{
global $modelSources, $serializedModelFile, $quranMetaDataFile, $META_DATA, $MODEL_CORE, $MODEL_SEARCH, $MODEL_QAC, $MODEL_QURANA;
global $UTHMANI_TO_SIMPLE_WORD_MAP, $numberOfSuras, $pauseMarksFile;
global $TRANSLATION_MAP_EN_TO_AR, $TRANSLATION_MAP_AR_TO_EN, $TRANSLITERATION_WORDS_MAP, $TRANSLITERATION_VERSES_MAP;
global $wordByWordTranslationFile, $transliterationFile;
global $MODEL_WORDNET, $qaOntologyNamespace, $qaOntologyFile, $is_a_relation_name_ar, $is_a_relation_name_en;
global $thing_class_name_ar, $thing_class_name_en;
global $MODEL_QA_ONTOLOGY, $arabicStopWordsFileL2;
global $TRANSLITERATION_WORDS_LOCATION_MAP, $TRANSLITERATION_WORDS_INDEX;
//not working
gc_enable();
if (!function_exists("apc_exists")) {
throw new Exception("APC not found!");
}
//echoN("MODEL EXISTS IN CACHE?:".apc_exists("EN/MODEL_CORE/TOTALS/"));
##### CHECK MODEL IN CACHE ##### #####
if (TRUE && apc_exists("EN/MODEL_CORE/TOTALS/") !== false) {
// split list by comma
$modelListArr = preg_split("/,/", trim($modelsToBeLoaded));
/**
* TODO: CHANGE THE CODE TO REFERENCE APC MEMORY DIRECTLY INSTEAD OF LOADING DATA IN EACH SCRIPT
*/
foreach ($modelListArr as $modelName) {
//echoN("$modelName $lang ".time());
// echoN(memory_get_peak_usage());
//echoN($modelName);
if ($modelName == "ontology") {
/*$MODEL_QA_ONTOLOGY = apc_fetch("MODEL_QA_ONTOLOGY");
if ($MODEL_QA_ONTOLOGY===false )
{
echo "$MODEL_QA_ONTOLOGY NOT CACHED";exit;
}
*/
}
if ($modelName == "wordnet") {
}
if ($modelName == "core") {
//$MODEL_CORE = json_decode((file_get_contents("$serializedModelFile.core")),TRUE);
/*$MODEL_CORE = apc_fetch("MODEL_CORE[$lang]");
if ($MODEL_CORE===false )
{
echo "CORE MODEL [$lang] NOT CACHED";exit;
}*/
} else {
if ($modelName == "search") {
//$MODEL_SEARCH = json_decode((file_get_contents("$serializedModelFile.search")),TRUE);
//$MODEL_SEARCH = apc_fetch("MODEL_SEARCH[$lang]");
/*if ($MODEL_SEARCH===false )
{
echo "SEARCH MODEL [$lang] NOT CACHED";exit;
}*/
} else {
if ($modelName == "qac") {
//$MODEL_QAC = json_decode((file_get_contents("$serializedModelFile.qac")),TRUE);
/*$MODEL_QAC = apc_fetch("MODEL_QAC");
if ($MODEL_QAC===false )
{
echo "QAC MODEL NOT CACHED";exit;
}
*/
}
}
}
}
$MODEL_WORDNET['INDEX'] = apc_fetch("WORDNET_INDEX");
if ($MODEL_WORDNET['INDEX'] === false) {
echo "MODEL_WORDNET['INDEX'] NOT CACHED";
exit;
}
$MODEL_WORDNET['LEXICO_SEMANTIC_CATEGORIES'] = apc_fetch("WORDNET_LEXICO_SEMANTIC_CATEGORIES");
if ($MODEL_WORDNET['LEXICO_SEMANTIC_CATEGORIES'] === false) {
echo " MODEL MODEL_WORDNET['LEXICO_SEMANTIC_CATEGORIES'] NOT CACHED";
exit;
}
$MODEL_WORDNET['DATA'] = apc_fetch("WORDNET_DATA");
if ($MODEL_WORDNET['DATA'] === false) {
echo "MODEL MODEL_WORDNET['DATA'] NOT CACHED";
exit;
}
//else if ( ($modelName=="qurana"))
//{
//$MODEL_QURANA = json_decode((file_get_contents("$serializedModelFile.qurana")),TRUE);
$MODEL_QURANA = apc_fetch("MODEL_QURANA");
if ($MODEL_QURANA === false) {
echo "QURANA MODEL NOT CACHED";
exit;
}
//}
return;
}
########## ##### ##### ##### ##### #####
//$quran = file($quranMetaDataFile);
$quranMetaDataXMLObj = simplexml_load_file($quranMetaDataFile);
###### CONVERT META XML STRUCUTURE TO OUR STRUCTURE
foreach ($quranMetaDataXMLObj->suras as $index => $surasArr) {
foreach ($surasArr->sura as $suraMetaArr) {
$tempArr = array();
$tempArr['index'] = (string) $suraMetaArr['index'];
$tempArr['ayas'] = (string) $suraMetaArr['ayas'];
$tempArr['name_ar'] = (string) $suraMetaArr['name'];
$tempArr['name_trans'] = (string) $suraMetaArr['tname'];
$tempArr['name_en'] = (string) $suraMetaArr['ename'];
$tempArr['type'] = (string) $suraMetaArr['type'];
$tempArr['order'] = (string) $suraMetaArr['order'];
$META_DATA['SURAS'][] = $tempArr;
}
}
##############################################
/////////// LOAD ONTOLOGY
$reader = new OWLReader();
$ontology = new OWLMemoryOntology();
$thingClassName = "{$thing_class_name_ar}";
$ontology->setNamespace($qaOntologyNamespace);
$reader->readFromFile($qaOntologyFile, $ontology);
//preprint_r($ontology->{'owl_data'}['classes']);
//preprint_r($ontology->{'owl_data'}['properties']);
//preprint_r($ontology->{'owl_data'}['labels']);
//preprint_r($ontology->{'owl_data'}['annotations']);
//preprint_r($ontology->{'owl_data'}['instances']);
$classes = $ontology->{'owl_data'}['classes'];
$instances = $ontology->{'owl_data'}['instances'];
$qaOntologyConceptsArr = array();
$qaOntologyRelationsArr = array();
$relationsCount = 0;
foreach ($classes as $className => $infoArr) {
$className = stripOntologyNamespace($className);
$qaOntologyConceptsArr[$className] = array("type" => "class");
//echoN($className);
//preprint_r($infoArr);
foreach ($infoArr[0]['properties'] as $index => $propertiesArr) {
/** INCASE THIS INSTANCE HAS MULTIPLE PROPERTIES WITH SAME VERB **/
foreach ($propertiesArr as $index2 => $onePropertyArr) {
if (empty($onePropertyArr)) {
continue;
}
$verb = key($onePropertyArr);
$objectClassArr = current($onePropertyArr);
$objectConceptName = stripOntologyNamespace($objectClassArr[0]);
//echoN("CLASS:***** $className => $verb -> $objectConceptName");
$attributedArr = next($onePropertyArr);
$freq = $attributedArr['frequency'];
$engTranslation = $attributedArr['verb_translation_en'];
$verbUthmani = $attributedArr['verb_uthmani'];
$relHashID = buildRelationHashID($className, $verb, $objectConceptName);
$qaOntologyRelationsArr[$relHashID] = array("SUBJECT" => $className, "VERB" => $verb, "OBJECT" => $objectConceptName, "FREQUENCY" => $freq, "VERB_TRANSLATION_EN" => $engTranslation, "VERB_UTHMANI" => $verbUthmani);
//preprint_r($qaOntologyRelationsArr[$relHashID]);
$relationsCount++;
}
}
}
foreach ($instances as $instanceName => $intancesArr) {
foreach ($intancesArr as $index => $infoArr) {
$subjectConceptName = stripOntologyNamespace($instanceName);
$parent = stripOntologyNamespace($infoArr['class']);
//echoN("$subjectConceptName $parent");
$relHashID = buildRelationHashID($subjectConceptName, $is_a_relation_name_ar, $parent);
$qaOntologyRelationsArr[$relHashID] = array("SUBJECT" => $subjectConceptName, "VERB" => "{$is_a_relation_name_ar}", "OBJECT" => $parent, "VERB_TRANSLATION_EN" => "{$is_a_relation_name_en}");
if ($parent != $thing_class_name_ar) {
$relationsCount++;
}
$propertiesArr = $infoArr['properties'];
//echoN($instanceName);
//echoN("$instanceName:@@@");
//preprint_r($propertiesArr);
/** INCASE THIS INSTANCE HAS MULTIPLE PROPERTIES WITH SAME VERB **/
foreach ($propertiesArr as $index2 => $onePropertyArr) {
if (empty($onePropertyArr)) {
continue;
}
$verb = key($onePropertyArr);
$objectClassArr = current($onePropertyArr);
$objectConceptName = stripOntologyNamespace($objectClassArr[0]);
//echoN("***** $verb -> $objectConceptName");
$attributedArr = next($onePropertyArr);
$freq = $attributedArr['frequency'];
$engTranslation = $attributedArr['verb_translation_en'];
$verbUthmani = $attributedArr['verb_uthmani'];
$relHashID = buildRelationHashID($subjectConceptName, $verb, $objectConceptName);
$qaOntologyRelationsArr[$relHashID] = array("SUBJECT" => $subjectConceptName, "VERB" => $verb, "OBJECT" => $objectConceptName, "FREQUENCY" => $freq, "VERB_TRANSLATION_EN" => $engTranslation, "VERB_UTHMANI" => $verbUthmani);
$relationsCount++;
}
// if it is class dont make it instance even if it is a subject (subclass of another class
// BUG: animal was not apearing on ontology graph page since it was instance
if (empty($qaOntologyConceptsArr[$subjectConceptName]) || $qaOntologyConceptsArr[$subjectConceptName][type] != 'class') {
$qaOntologyConceptsArr[$subjectConceptName] = array("type" => "instance");
}
}
}
foreach ($qaOntologyConceptsArr as $conceptName => $infoArr) {
$fullConceptName = $qaOntologyNamespace . $conceptName;
$labelsArr = $ontology->{'owl_data'}['labels'][$fullConceptName];
foreach ($labelsArr as $labelLang => $label) {
/*if ( mb_strlen($label)==1)
{
echon($fullConceptName);
preprint_r($ontology->{'owl_data'}['labels'][$fullConceptName]);
}*/
$qaOntologyConceptsArr[$conceptName]['label_' . strtolower($labelLang)] = $label;
}
// "Thing" does not have annotations
if (isset($ontology->{'owl_data'}['annotations'][$fullConceptName])) {
$annotationsArr = $ontology->{'owl_data'}['annotations'][$fullConceptName];
foreach ($annotationsArr as $index => $annotArr) {
$key = $annotArr['KEY'];
$val = $annotArr['VAL'];
$qaOntologyConceptsArr[$conceptName][$key] = $val;
//echoN("[$conceptName][$key] = $val");
}
}
}
////////// OUTPUT STATS
/*echoN("INSTANCES COUNT:".count($ontology->{'owl_data'}['instances']));
echoN("CLASSES COUNT:".count($ontology->{'owl_data'}['classes']));
echoN("PROPERTIES COUNT - DECLERATIONS ONLY:".count($ontology->{'owl_data'}['properties']));;
echoN("CONCEPTS COUNT:".count($qaOntologyConceptsArr));
echoN("RELATIONS COUNT:".$relationsCount);
preprint_r($qaOntologyRelationsArr);*/
//////////////////
///////////// QUALITY CHECK CONCEPTS
$qaOntologyConceptsArr2 = array();
foreach ($qaOntologyConceptsArr as $key => $val) {
$newKey = strtr($key, "_", " ");
$qaOntologyConceptsArr2[$newKey] = $value;
}
$ONTOLOGY_EXTRACTION_FOLDER = "../data/ontology/extraction/";
$finalConcepts = unserialize(file_get_contents("{$ONTOLOGY_EXTRACTION_FOLDER}/temp.final.concepts.final"));
$diffArr = array_diff(array_keys($qaOntologyConceptsArr2), array_keys($finalConcepts));
$conceptsDiffCount = count($matchingTable);
if ($relationsDiffCount > 0) {
echoN("<b>### OWL-PROPRIETARY-CONCEPTS-DIFF-COUNT:</b>" . $conceptsDiffCount);
}
//preprint_r($diffArr);
//////////////////////////////////////////////////////////////
//////// quality check relations
$relationsArr = unserialize(file_get_contents("{$ONTOLOGY_EXTRACTION_FOLDER}/temp.final.relations"));
$matchingTable = array();
foreach ($qaOntologyRelationsArr as $index => $relArr) {
$trippleStr = $relArr['SUBJECT'] . "->" . $relArr['VERB'] . "->" . $relArr['OBJECT'];
//since Thing relations are not in the list we are comparing with
if ($relArr['OBJECT'] == $thing_class_name_ar) {
continue;
}
//echoN($trippleStr);
$trippleStr = trim($trippleStr);
$matchingTable[$trippleStr]++;
}
foreach ($relationsArr as $index => $relArr) {
$relArr['SUBJECT'] = strtr($relArr['SUBJECT'], " ", "_");
$relArr['VERB'] = strtr($relArr['VERB'], " ", "_");
$relArr['OBJECT'] = strtr($relArr['OBJECT'], " ", "_");
$trippleStr = $relArr['SUBJECT'] . "->" . $relArr['VERB'] . "->" . $relArr['OBJECT'];
$trippleStr = trim($trippleStr);
$matchingTable[$trippleStr]++;
}
function filterFunc($v)
{
return $v <= 1;
}
$matchingTable = array_filter($matchingTable, 'filterFunc');
$relationsDiffCount = count($matchingTable);
if ($relationsDiffCount > 0) {
echoN("<b>### OWL-PROPRIETARY-RELATIONS-DIFF-COUNT:</b>" . $relationsDiffCount);
preprint_r($matchingTable);
}
//////////////////////////////////////////////
//echoN( join("<br>",array_keys($qaOntologyConceptsArr)));
$qaOntologyVerbIndex = array();
$qaOntologyGraphSourcesIndex = array();
$qaOntologyGraphTargetsIndex = array();
//preprint_r($qaOntologyRelationsArr);
//exit;
foreach ($qaOntologyRelationsArr as $index => $relArr) {
$subject = $relArr['SUBJECT'];
$verb = $relArr['VERB'];
$verb_translation_en = $relArr['VERB_TRANSLATION_EN'];
$object = $relArr['OBJECT'];
//$qaOntologyVerbIndex[$verb][]=array("SUBJECT"=>$subject,"OBJECT"=>$object);
//$qaOntologyVerbIndex[$verb_translation_en][]=array("SUBJECT"=>$subject,"OBJECT"=>$object);
addValueToMemoryModel("ALL", "MODEL_QA_ONTOLOGY", "VERB_INDEX", $verb, array("SUBJECT" => $subject, "OBJECT" => $object));
addValueToMemoryModel("ALL", "MODEL_QA_ONTOLOGY", "VERB_INDEX", $verb_translation_en, array("SUBJECT" => $subject, "OBJECT" => $object));
//$qaOntologyGraphSourcesIndex[$subject][]=array("link_verb"=>$verb,"target"=>$object,"relation_index"=>$index);
//$qaOntologyGraphTargetsIndex[$object][]=array("source"=>$subject,"link_verb"=>$verb,"relation_index"=>$index);
addToMemoryModelList("ALL", "MODEL_QA_ONTOLOGY", "GRAPH_INDEX_SOURCES", $subject, array("link_verb" => $verb, "target" => $object, "relation_index" => $index));
addToMemoryModelList("ALL", "MODEL_QA_ONTOLOGY", "GRAPH_INDEX_TARGETS", $object, array("source" => $subject, "link_verb" => $verb, "relation_index" => $index));
}
$qaOntologyConceptsENtoARMapArr = array();
foreach ($qaOntologyConceptsArr as $arName => $conceptArr) {
$enLabel = trim(strtolower($conceptArr['label_en']));
//$qaOntologyConceptsENtoARMapArr[$enLabel]=$arName;
//$qaOntologyConceptsENtoARMapArr[$enLabel]=$arName;
addValueToMemoryModel("ALL", "MODEL_QA_ONTOLOGY", "CONCEPTS_EN_AR_NAME_MAP", $enLabel, $arName);
}
$qaSynonymsIndex = array();
foreach ($qaOntologyConceptsArr as $arName => $conceptArr) {
addValueToMemoryModel("ALL", "MODEL_QA_ONTOLOGY", "CONCEPTS", $arName, $conceptArr);
$i = 1;
while (isset($conceptArr['synonym_' . $i])) {
if (empty($conceptArr['synonym_' . $i])) {
$i++;
continue;
}
$synonymLabel = trim(strtolower($conceptArr['synonym_' . $i]));
$qaSynonymsIndex[$synonymLabel] = $arName;
addValueToMemoryModel("ALL", "MODEL_QA_ONTOLOGY", "SYNONYMS_INDEX", $synonymLabel, $arName);
$i++;
}
}
//preprint_r($qaOntologyConceptsArr);exit;
//$MODEL_QA_ONTOLOGY['CONCEPTS'] = $qaOntologyConceptsArr;
//$MODEL_QA_ONTOLOGY['RELATIONS'] = $qaOntologyRelationsArr;
addValueToMemoryModel("ALL", "MODEL_QA_ONTOLOGY", "RELATIONS", "", $qaOntologyRelationsArr);
//$MODEL_QA_ONTOLOGY['GRAPH_INDEX_SOURCES'] = $qaOntologyGraphSourcesIndex;
//$MODEL_QA_ONTOLOGY['GRAPH_INDEX_TARGETS'] = $qaOntologyGraphTargetsIndex;
//$MODEL_QA_ONTOLOGY['CONCEPTS_EN_AR_NAME_MAP'] = $qaOntologyConceptsENtoARMapArr;
//$MODEL_QA_ONTOLOGY['VERB_INDEX'] = $qaOntologyVerbIndex;
//$MODEL_QA_ONTOLOGY['SYNONYMS_INDEX'] = $qaSynonymsIndex;
//$res = apc_store("MODEL_QA_ONTOLOGY",$MODEL_QA_ONTOLOGY);
//if ( $res===false){ throw new Exception("Can't cache MODEL_QA_ONTOLOGY"); }
//preprint_r($MODEL_QA_ONTOLOGY);exit;
//////// END ONTOLOGY LOADING
////////////////////////////
/// WORDNET
loadWordnet($MODEL_WORDNET);
/////////////
//free resources
$quranMetaDataXMLObj = null;
unset($quranMetaDataXMLObj);
foreach ($modelSources as $supportedLang => $modelSourceArr) {
$type = $modelSourceArr['type'];
$file = $modelSourceArr['file'];
//echoN("$lang $type $file");
loadModel($supportedLang, $type, $file);
//not working
$gced = gc_collect_cycles();
//echoN($gced);
}
//echoN(json_encode($MODEL));
############ Uthmani/Simple mapping table #################
############ AND WORD-WORD TRANSLATION AND TRANSLITERATION #################
$pauseMarksArr = getPauseMarksArrByFile($pauseMarksFile);
$wordByWordFileArr = file($wordByWordTranslationFile, FILE_SKIP_EMPTY_LINES | FILE_IGNORE_NEW_LINES);
$translitertationArr = file($transliterationFile, FILE_SKIP_EMPTY_LINES | FILE_IGNORE_NEW_LINES);
$WORD_SENSES_EN = array();
$WORD_SENSES_AR = array();
$quranTextEntryFromAPC_AR = getModelEntryFromMemory("AR", "MODEL_CORE", "QURAN_TEXT", "");
$quranTextEntryFromAPC_UTH = getModelEntryFromMemory("AR_UTH", "MODEL_CORE", "QURAN_TEXT", "");
/* SURA'S LOOP **/
for ($s = 0; $s < $numberOfSuras; $s++) {
$suraSize = count($quranTextEntryFromAPC_AR[$s]);
/* VERSES LOOP **/
for ($a = 0; $a < $suraSize; $a++) {
$i++;
$verseTextSimple = $quranTextEntryFromAPC_AR[$s][$a];
$simpleWordsArr = preg_split("/ /", $verseTextSimple);
$verseTextUthmani = $quranTextEntryFromAPC_UTH[$s][$a];
$uthmaniWordsArr = preg_split("/ /", $verseTextUthmani);
$simpleWordsArr = removePauseMarksFromArr($pauseMarksArr, $simpleWordsArr);
$uthmaniWordsArr = removePauseMarksFromArr($pauseMarksArr, $uthmaniWordsArr);
$verseLocation = $s + 1 . ":" . ($a + 1);
$UTHMANI_TO_SIMPLE_LOCATION_MAP[$verseLocation] = array();
///////// Transliteration /////////////
$transliterationLine = current($translitertationArr);
next($translitertationArr);
$lineParts = preg_split("/\\|/", $transliterationLine);
$verseTransliteration = $lineParts[2];
//echoN($transliterationLine);
$TRANSLITERATION_VERSES_MAP[$verseLocation] = $verseTransliteration;
$wordsTransliterationArr = preg_split("/ /", $verseTransliteration);
// preprint_r($wordsTransliterationArr);exit;
/////////////////////////////////////////////////
$wtwIndex = 0;
foreach ($uthmaniWordsArr as $index => $wordUthmani) {
$qacMasterID = $s + 1 . ":" . ($a + 1) . ":" . ($index + 1);
$qacMasterTableEntry = getModelEntryFromMemory("AR", "MODEL_QAC", "QAC_MASTERTABLE", $qacMasterID);
$lemma = $qacMasterTableEntry[0]['FEATURES']['LEM'];
// to handle multi segment words such as الدنيا
if (empty($lemma)) {
$lemma = $qacMasterTableEntry[1]['FEATURES']['LEM'];
}
//echoN("|$lemma|$wordUthmani");
//$wtwIndex (INDEX_IN_AYA_EMLA2Y) needs to be 1 based ( UTHMANI=IMLA2Y )
$UTHMANI_TO_SIMPLE_LOCATION_MAP[$s + 1 . ":" . ($a + 1)][$index + 1] = $wtwIndex + 1;
$wordSimple = $simpleWordsArr[$wtwIndex++];
//$UTHMANI_TO_SIMPLE_LOCATION_MAP[($s+1).":".($a+1)][($index+1)."-".$wordUthmani]=($wtwIndex)."-".$wordSimple;
/* for ayas which are different in size, do the following
* if the current word is ويا or ها or يا
* then join it with the next word and make them one word
*/
if (count($uthmaniWordsArr) != count($simpleWordsArr) && ($wordSimple == "يا" || $wordSimple == "ها" || $wordSimple == "ويا" || $wordUthmani == "وَأَلَّوِ")) {
if ($wordUthmani == "يَبْنَؤُمَّ") {
// example 0 => 1
$UTHMANI_TO_SIMPLE_LOCATION_MAP[$s + 1 . ":" . ($a + 1)][$index + 1] = $wtwIndex + 1;
//[($index+1)."-".$wordUthmani]=($wtwIndex+1)."-".$wordSimple;
$wordSimple = $wordSimple . " " . $simpleWordsArr[$wtwIndex++] . " " . $simpleWordsArr[$wtwIndex++];
} else {
// example 0 => 1
$UTHMANI_TO_SIMPLE_LOCATION_MAP[$s + 1 . ":" . ($a + 1)][$index + 1] = $wtwIndex + 1;
//[($index+1)."-".$wordUthmani]=($wtwIndex+1)."-".$wordSimple;
$wordSimple = $wordSimple . " " . $simpleWordsArr[$wtwIndex++];
}
//echoN("$wordUthmani:$wordSimple");
}
// printHTMLPageHeader();
// echoN("$wordSimple|$wordUthmani");
///////// english translation ////////
$wordByWordTranslationLine = current($wordByWordFileArr);
next($wordByWordFileArr);
$linePartsArr = preg_split("/\\|/", $wordByWordTranslationLine);
$englishTranslationForCurrentWord = $linePartsArr[5];
/////////////////////////////////////////////////
$WORD_SENSES_EN[$englishTranslationForCurrentWord][$wordUthmani]++;
$WORD_SENSES_AR[$wordUthmani][$englishTranslationForCurrentWord]++;
$TRANSLATION_MAP_EN_TO_AR[$englishTranslationForCurrentWord] = $wordUthmani;
$TRANSLATION_MAP_AR_TO_EN[$wordUthmani] = $englishTranslationForCurrentWord;
$TRANSLITERATION_WORDS_MAP[$wordUthmani] = $wordsTransliterationArr[$index];
$clenaedTranliteration = cleanTransliteratedText($wordsTransliterationArr[$index]);
$TRANSLITERATION_WORDS_INDEX[$clenaedTranliteration] = 1;
$TRANSLITERATION_WORDS_LOCATION_MAP["{$s}:{$a}:{$index}"] = $wordsTransliterationArr[$index];
//preprint_r($TRANSLITERATION_WORDS_LOCATION_MAP);
// preprint_r($TRANSLATION_MAP_AR_TO_EN);
// preprint_r($TRANSLITERATION_WORDS_MAP);
$UTHMANI_TO_SIMPLE_WORD_MAP[$wordUthmani] = $wordSimple;
addValueToMemoryModel("AR", "OTHERS", "UTHMANI_TO_SIMPLE_WORD_MAP", $wordUthmani, $wordSimple);
$UTHMANI_TO_SIMPLE_WORD_MAP[$wordSimple] = $wordUthmani;
addValueToMemoryModel("AR", "OTHERS", "UTHMANI_TO_SIMPLE_WORD_MAP", $wordSimple, $wordUthmani);
if (!empty($lemma)) {
if (!isset($LEMMA_TO_SIMPLE_WORD_MAP[$lemma])) {
$LEMMA_TO_SIMPLE_WORD_MAP[$lemma] = $wordSimple;
} else {
$oldSimple = $LEMMA_TO_SIMPLE_WORD_MAP[$lemma];
if (myLevensteinEditDistance($oldSimple, $lemma) > myLevensteinEditDistance($wordSimple, $lemma)) {
$LEMMA_TO_SIMPLE_WORD_MAP[$lemma] = $wordSimple;
}
}
}
}
}
}
/////// ADD UTHMANI TO SIMPLE LOCATION MAP TO MEMORY
foreach ($UTHMANI_TO_SIMPLE_LOCATION_MAP as $verseLocation => $verseMappingArr) {
/*foreach($mappingArr as $uhtmaniIndex=>$imal2yIndex)
{
}*/
addValueToMemoryModel("AR", "OTHERS", "UTHMANI_TO_SIMPLE_LOCATION_MAP", $verseLocation, $verseMappingArr);
}
///////////////////////////////////////////////////////
//preprint_r($TRANSLATION_MAP_EN_TO_AR);exit;
//preprint_r($WORD_SENSES_AR);exit;
// CAN'T BE ADDED IN THE CORE_MODEL since the mapping happens after loadModel
//$res = apc_store("UTHMANI_TO_SIMPLE_WORD_MAP",$UTHMANI_TO_SIMPLE_WORD_MAP);
//if ( $res===false){ throw new Exception("Can't cache UTHMANI_TO_SIMPLE_WORD_MAP"); }
//$res = apc_store("UTHMANI_TO_SIMPLE_LOCATION_MAP",$UTHMANI_TO_SIMPLE_LOCATION_MAP);
//if ( $res===false){ throw new Exception("Can't cache UTHMANI_TO_SIMPLE_LOCATION_MAP"); }
$res = apc_store("LEMMA_TO_SIMPLE_WORD_MAP", $LEMMA_TO_SIMPLE_WORD_MAP);
if ($res === false) {
throw new Exception("Can't cache LEMMA_TO_SIMPLE_WORD_MAP");
}
$res = apc_store("WORDS_TRANSLATIONS_EN_AR", $TRANSLATION_MAP_EN_TO_AR);
if ($res === false) {
throw new Exception("Can't cache WORDS_TRANSLATIONS_EN_AR");
}
$res = apc_store("WORDS_TRANSLATIONS_AR_EN", $TRANSLATION_MAP_AR_TO_EN);
if ($res === false) {
throw new Exception("Can't cache WORDS_TRANSLATIONS_AR_EN");
}
$res = apc_store("WORDS_TRANSLITERATION", $TRANSLITERATION_WORDS_MAP);
if ($res === false) {
throw new Exception("Can't cache WORDS_TRANSLITERATION");
}
$res = apc_store("TRANSLITERATION_WORDS_LOCATION_MAP", $TRANSLITERATION_WORDS_LOCATION_MAP);
if ($res === false) {
throw new Exception("Can't cache TRANSLITERATION_WORDS_LOCATION_MAP");
}
$res = apc_store("TRANSLITERATION_VERSES_MAP", $TRANSLITERATION_VERSES_MAP);
if ($res === false) {
throw new Exception("Can't cache TRANSLITERATION_VERSES_MAP");
}
$res = apc_store("TRANSLITERATION_WORDS_INDEX", $TRANSLITERATION_WORDS_INDEX);
if ($res === false) {
throw new Exception("Can't cache TRANSLITERATION_WORDS_INDEX");
}
$res = apc_store("WORD_SENSES_EN", $WORD_SENSES_EN);
if ($res === false) {
throw new Exception("Can't cache WORD_SENSES_EN");
}
$res = apc_store("WORD_SENSES_AR", $WORD_SENSES_AR);
if ($res === false) {
throw new Exception("Can't cache {$WORD_SENSES_AR}");
}
//// ENRICH INVERTED INDEX BY UTHMANI-EMLA2Y INDEXES
//echoN(count($MODEL_SEARCH['AR']['INVERTED_INDEX']));
foreach (getAPCIterator("AR\\/MODEL_SEARCH\\/INVERTED_INDEX\\/.*") as $invertedIndexCursor) {
$wordDataArr = $invertedIndexCursor['value'];
$key = $invertedIndexCursor['key'];
$word = getEntryKeyFromAPCKey($key);
foreach ($wordDataArr as $index => $documentArrInIndex) {
$WORD_TYPE = $documentArrInIndex['WORD_TYPE'];
$SURA = $documentArrInIndex['SURA'];
$AYA = $documentArrInIndex['AYA'];
//echoN($word." ".$WORD_TYPE);
if ($WORD_TYPE == "NORMAL_WORD") {
$INDEX_IN_AYA_EMLA2Y = $documentArrInIndex['INDEX_IN_AYA_EMLA2Y'];
foreach ($UTHMANI_TO_SIMPLE_LOCATION_MAP[$SURA + 1 . ":" . ($AYA + 1)] as $uhtmaniIndex => $imal2yIndex) {
if ($imal2yIndex == $INDEX_IN_AYA_EMLA2Y) {
$INDEX_IN_AYA_UTHMANI = $uhtmaniIndex;
break;
}
}
//echoN($INDEX_IN_AYA_UTHMANI);
$wordDataArr[$index]['INDEX_IN_AYA_UTHMANI'] = $INDEX_IN_AYA_UTHMANI;
} else {
// needed for highlighting pronoun charcters in search
$INDEX_IN_AYA_UTHMANI = $documentArrInIndex['INDEX_IN_AYA_UTHMANI'];
$INDEX_IN_AYA_EMLA2Y = getSimpleWordIndexByUthmaniWordIndex($SURA + 1 . ":" . ($AYA + 1), $INDEX_IN_AYA_UTHMANI);
$wordDataArr[$index]['INDEX_IN_AYA_EMLA2Y'] = $INDEX_IN_AYA_EMLA2Y;
}
}
//UPDATE
updateModelData($key, $wordDataArr);
}
//$res = apc_store("MODEL_SEARCH[AR]",$MODEL_SEARCH['AR']);
//if ( $res===false){ throw new Exception("Can't cache MODEL_SEARCH[AR]"); }
//preprint_r($TRANSLITERATION_WORDS_LOCATION_MAP);
/// ADD TRANSLITERATION TO INVERETD INDEX WWITH ENGLISH WORDS
if ($lang == "EN") {
$invertedIndexBatchApcArr = array();
foreach ($TRANSLITERATION_WORDS_LOCATION_MAP as $location => $transliteratedWord) {
$locationArr = explode(":", $location);
$s = $locationArr[0];
$a = $locationArr[1];
$wordIndex = $locationArr[2];
//echoN("$transliteratedWord,$s,$a,$wordIndex");
$transliteratedWord = strtolower(strip_tags($transliteratedWord));
//$MODEL_SEARCH['EN']['INVERTED_INDEX'][$word]
addToInvertedIndex($invertedIndexBatchApcArr, $lang, $transliteratedWord, $s, $a, $wordIndex, "NORMAL_WORD");
}
addToMemoryModelBatch($invertedIndexBatchApcArr);
//$res = apc_store("MODEL_SEARCH[EN]",$MODEL_SEARCH['EN']);
}
//if ( $res===false){ throw new Exception("Can't cache MODEL_SEARCH[EN]"); }
/////////////////////////////////////////////////////////
//preprint_r($UTHMANI_TO_SIMPLE_WORD_MAP);
//preprint_r($MODEL_CORE["AR_UTH"]['QURAN_TEXT']);exit;
##############################################################
// get memory usage
$debug = memory_get_usage(true) / 1024 / 1024 . "/" . memory_get_peak_usage(true) / 1024 / 1024 . "Memory <br>";
//echoN($debug);
//needed to be set here after both languages has been loaded
// reload all models from memory to set all variables (WORDNET) - after model generation
/* needed to reload all generated models from memory specialy model_core since
* it has 3 languages, if this line is removed: all 3 langauges are loaded although only one language
* is requested, also it caused a bug in getPoSTaggedSubsentences
*/
//loadModels($modelsToBeLoaded,$lang);
}
function extendQueryWordsByDerivations($taggedSignificantWords, $lang)
{
foreach ($taggedSignificantWords as $word => $posTag) {
// avoid small words, will lead to too many iirelevant derivations
if (mb_strlen($word) <= 2) {
continue;
}
if ($lang == "EN") {
if ($posTag == "NN") {
$plural = $word . "s";
$taggedSignificantWords[$plural] = "{$posTag}" . "S";
} else {
if ($posTag == "NNS") {
$single = substr($word, 0, -1);
$taggedSignificantWords[$single] = substr($posTag, 0, -1);
}
}
} else {
$simmlarWords = array();
$qaOntologyConceptsIterator = getAPCIterator("ALL\\/MODEL_QA_ONTOLOGY\\/CONCEPTS\\/.*");
foreach ($qaOntologyConceptsIterator as $conceptsCursor) {
$conceptID = getEntryKeyFromAPCKey($conceptsCursor['key']);
$mainConceptArr = $conceptsCursor['value'];
$conceptLabelAR = $mainConceptArr['label_ar'];
$dist = myLevensteinEditDistance($word, $conceptLabelAR);
if ($dist <= 5) {
$dist = getDistanceByCommonUniqueChars($word, $conceptLabelAR);
$simmlarWords[$conceptLabelAR] = $dist;
}
$i = 1;
while (isset($mainConceptArr['synonym_' . $i]) && isArabicString($mainConceptArr['synonym_' . $i])) {
$synonym = $mainConceptArr['synonym_' . $i];
$dist = myLevensteinEditDistance($word, $synonym);
if ($dist <= 5) {
$dist = getDistanceByCommonUniqueChars($word, $synonym);
$simmlarWords[$synonym] = $dist;
}
$i++;
}
}
foreach ($simmlarWords as $conceptWord => $distBySimChars) {
$diff = mb_strlen($conceptWord) - mb_strlen($word);
$absDiffSize = abs($diff);
// $word is bigger
if ($diff < 0) {
$absDiffSize = abs($diff);
$diffStr = getStringDiff($conceptWord, $word);
//echoN($diffStr);
//حيوان => الحيوانات
// the bigger word should not contain space الله => سبيل الله
// $diffStr=="الات" for حيوان = الحياوانات
if (mb_strpos($word, $conceptWord) !== false && strpos($word, " ") === false && ($diffStr == "ات" || $diffStr == "ال" || $diffStr == "الات")) {
//echoN("$word, $conceptWord");
/// convert word to noun plular
$taggedSignificantWords[$word] = "NNS";
// concept word is singular
$taggedSignificantWords[$conceptWord] = "NN";
} else {
if ($diff == 1) {
$wordLastCharTrimmed = mb_substr($conceptWord, 0, -1);
if ($wordLastCharTrimmed . "ات" == $word) {
/// convert word to noun plular
$taggedSignificantWords[$word] = "NNS";
/// convert word to noun plular
$taggedSignificantWords[$conceptWord] = "NN";
}
}
}
} else {
$diffStr = getStringDiff($conceptWord, $word);
// الحيوانات => حيوان
// the bigger word should not contain space الله => سبيل الله
if ($diff != 0 && mb_strpos($conceptWord, $word) !== false && strpos($conceptWord, " ") === false && ($diffStr == "ات" || $diffStr == "ال" || $diffStr == "الات")) {
//echoN("$word,$conceptWord");
/// convert word to noun plular
$taggedSignificantWords[$conceptWord] = "NNS";
} else {
if ($diff == 1) {
$wordLastCharTrimmed = mb_substr($word, 0, -1);
if ($wordLastCharTrimmed . "ات" == $conceptWord) {
/// convert word to noun plular
$taggedSignificantWords[$conceptWord] = "NNS";
}
}
}
}
}
// limit the number of derivations+original terms to 15
$taggedSignificantWords = array_slice($taggedSignificantWords, 0, 10);
//arsort($simmlarWords);
//preprint_r($taggedSignificantWords);
//preprint_r($simmlarWords);
//exit;
}
}
//preprint_r($taggedSignificantWords);
return $taggedSignificantWords;
}
