function createLexemDefinitionMap()
{
LexemDefinitionMap::deleteAll();
$dbResult = db_selectAllConcepts();
print "Migrating " . mysql_num_rows($dbResult) . " concepts...\n";
$seen = 0;
while ($dbRow = mysql_fetch_assoc($dbResult)) {
$concept = new Concept();
$concept->populateFromDbRow($dbRow);
$words = Word::loadByConceptId($concept->id);
$definitions = Definition::loadByConceptId($concept->id);
if ($definitions) {
// Select distinct words
$distinctWords = array();
foreach ($words as $word) {
$distinctWords[$word->name] = 1;
}
// For every word, look up all the lexems. Then map each of those lexems
// to every definition.
foreach ($distinctWords as $word => $ignored) {
$lexems = Lexem::loadByUnaccented($word);
// Create lexem if necessary so that we don't lose any words during the
// migration
if (count($lexems) == 0) {
$lexem = Lexem::create($word, 'T', 1, '');
$lexem->save();
$lexem->id = db_getLastInsertedId();
$lexems[] = $lexem;
$lexem->regenerateParadigm();
}
foreach ($lexems as $lexem) {
foreach ($definitions as $definition) {
$ldm = LexemDefinitionMap::load($lexem->id, $definition->id);
if (!$ldm) {
$ldm = LexemDefinitionMap::create($lexem->id, $definition->id);
$ldm->save();
}
}
}
}
}
$seen++;
if ($seen % 1000 == 0) {
print "Seen: {$seen};\n";
}
}
print "Seen: {$seen};\n";
}
continue;
}
$firstAt = strpos($def->internalRep, '@');
$secondAt = strpos($def->internalRep, '@', $firstAt + 1);
assert($firstAt === 0);
assert($secondAt !== false);
$text = trim(substr($def->internalRep, $firstAt + 1, $secondAt - $firstAt - 1));
$normText = text_removeAccents(text_unicodeToLower($text));
if (text_contains($normText, '-') && str_replace('-', '', $normText) == $l->unaccented) {
print "[{$l->unaccented}] [{$normText}]\n";
$parts = split('-', $normText);
foreach ($parts as $part) {
$lexems = Lexem::searchWordlists($part, true);
if (!count($lexems)) {
print "Creez lexemul [{$part}]\n";
$lexem = Lexem::create($part, 'T', '1', '');
$lexem->comment = 'Creat pentru despărțirea în cuvinte a unui ' . 'alt lexem';
$lexem->save();
$lexem->id = db_getLastInsertedId();
$lexem->regenerateParadigm();
$lexems[] = $lexem;
}
// Now associate every lexem with every definition
foreach ($defs as $defAssoc) {
foreach ($lexems as $lexemAssoc) {
LexemDefinitionMap::associate($lexemAssoc->id, $defAssoc->id);
}
}
}
foreach ($defs as $fixDef) {
$fixFirstAt = strpos($fixDef->internalRep, '@');
$definition->abbrevReview = count($ambiguousMatches) ? ABBREV_AMBIGUOUS : ABBREV_REVIEW_COMPLETE;
$definition->save();
log_userLog("Added definition {$definition->id} ({$definition->lexicon})");
$ldms = array();
foreach ($lexemNames as $lexemName) {
$lexemName = addslashes(AdminStringUtil::formatLexem($lexemName));
if ($lexemName) {
$matches = Lexem::loadByExtendedName($lexemName);
if (count($matches) >= 1) {
foreach ($matches as $match) {
LexemDefinitionMap::associate($match->id, $definition->id);
log_userLog("Associating with lexem {$match->id} ({$match->form})");
}
} else {
// Create a new lexem.
$lexem = Lexem::create($lexemName, 'T', '1', '');
$lexem->save();
$lexem->regenerateParadigm();
LexemDefinitionMap::associate($lexem->id, $definition->id);
log_userLog("Created lexem {$lexem->id} ({$lexem->form})");
}
}
}
FlashMessage::add('Definiția a fost trimisă. Un moderator o va examina în scurt timp. Vă mulțumim!', 'info');
util_redirect('contribuie');
}
} else {
smarty_assign('sourceId', session_getDefaultContribSourceId());
}
smarty_assign('contribSources', Model::factory('Source')->where('canContribute', true)->order_by_asc('displayOrder')->find_many());
smarty_assign('page_title', 'Contribuie cu definiții');
smarty_assign('selectedModelNumber', $modelNumber);
LocVersion::changeDatabase($locVersion);
if ($modelNumber == -1) {
$modelsToDisplay = FlexModel::loadByType($modelType);
} else {
$modelsToDisplay = array(Model::factory('FlexModel')->where('modelType', $modelType)->where('number', $modelNumber)->find_one());
}
$lexems = array();
$paradigms = array();
foreach ($modelsToDisplay as $m) {
// Load by canonical model, so if $modelType is V, look for a lexem with type V or VT.
$l = Model::factory('Lexem')->select('Lexem.*')->join('ModelType', 'modelType = code')->where('canonical', $modelType)->where('modelNumber', $m->number)->where('form', $m->exponent)->limit(1)->find_one();
if ($l) {
$paradigm = getExistingForms($l->id, $locVersion);
} else {
$l = Lexem::create($m->exponent, $modelType, $m->number, '');
$l->isLoc = true;
$paradigm = getNewForms($l, $locVersion);
}
$lexems[] = $l;
$paradigms[] = $paradigm;
}
smarty_assign('modelsToDisplay', $modelsToDisplay);
smarty_assign('lexems', $lexems);
smarty_assign('paradigms', $paradigms);
} else {
smarty_assign('selectedLocVersion', $locVersions[0]->name);
// LocVersion::changeDatabase($locVersion);
}
$modelTypes = ModelType::loadCanonical();
$models = FlexModel::loadByType($modelType ? $modelType : $modelTypes[0]->code);
require_once "../../phplib/util.php";
ini_set('memory_limit', '512M');
ini_set('max_execution_time', '3600');
util_assertModerator(PRIV_LOC);
util_assertNotMirror();
DebugInfo::disable();
$modelType = util_getRequestParameter('modelType');
$modelNumber = util_getRequestParameter('modelNumber');
$previewButton = util_getRequestParameter('previewButton');
$confirmButton = util_getRequestParameter('confirmButton');
$modelType = ModelType::canonicalize($modelType);
$inflections = Model::factory('Inflection')->where('modelType', $modelType)->order_by_asc('rank')->find_many();
// Load the original data
$model = Model::factory('FlexModel')->where('modelType', $modelType)->where('number', $modelNumber)->find_one();
$exponent = $model->exponent;
$lexem = Lexem::create($exponent, $modelType, $modelNumber, '');
$ifs = $lexem->generateParadigm();
$mdMap = ModelDescription::getByModelIdMapByInflectionIdVariantApplOrder($model->id);
$forms = array();
foreach ($inflections as $infl) {
$forms[$infl->id] = array();
}
foreach ($ifs as $if) {
$forms[$if->inflectionId][] = array('form' => $if->form, 'isLoc' => $mdMap[$if->inflectionId][$if->variant][0]->isLoc, 'recommended' => $mdMap[$if->inflectionId][$if->variant][0]->recommended);
}
$participleNumber = $modelType == 'V' ? ParticipleModel::loadByVerbModel($modelNumber)->adjectiveModel : '';
if ($previewButton || $confirmButton) {
// Load the new forms and exponent;
$newModelNumber = util_getRequestParameter('newModelNumber');
$newExponent = util_getRequestParameter('newExponent');
$newDescription = util_getRequestParameter('newDescription');
function _cloneLexem($lexem)
{
$clone = Lexem::create($lexem->form, 'T', 1, '');
$clone->comment = $lexem->comment;
$clone->description = $lexem->description ? "CLONĂ {$lexem->description}" : "CLONĂ";
$clone->noAccent = $lexem->noAccent;
$clone->save();
// Clone the definition list
$ldms = LexemDefinitionMap::get_all_by_lexemId($lexem->id);
foreach ($ldms as $ldm) {
LexemDefinitionMap::associate($clone->id, $ldm->definitionId);
}
$clone->regenerateParadigm();
return $clone;
}
foreach ($lexemNames as $lexemName) {
$lexemName = trim($lexemName);
if ($lexemName) {
$matches = Lexem::loadByExtendedName($lexemName);
if (count($matches) >= 1) {
foreach ($matches as $match) {
if (!in_array($match->id, $lexemIds)) {
$lexemIds[] = $match->id;
$lexems[] = $match;
$ldms[] = LexemDefinitionMap::create($match->id, $definitionId);
}
}
} else {
$hasErrors = true;
FlashMessage::add("Lexemul <i>" . htmlentities($lexemName) . "</i> nu există. Folosiți lista de sugestii pentru a-l corecta.");
$lexems[] = Lexem::create($lexemName, 0, '', '');
// We won't be needing $ldms since there is an error.
}
}
}
} else {
$lexems = Model::factory('Lexem')->select('Lexem.*')->join('LexemDefinitionMap', 'Lexem.id = lexemId')->where('definitionId', $definitionId)->find_many();
}
if ($commentContents) {
if (!$comment) {
$comment = Model::factory('comment')->create();
$commend->status = ST_ACTIVE;
$comment->definitionId = $definitionId;
}
$newContents = AdminStringUtil::internalizeDefinition($commentContents, $sourceId);
if ($newContents != $comment->contents) {
public function regenerateLongInfinitive()
{
$infl = Inflection::loadLongInfinitive();
$ifs = Model::factory('InflectedForm')->where('lexemId', $this->id)->where('inflectionId', $infl->id)->find_many();
$f107 = Model::factory('FlexModel')->where('modelType', 'F')->where('number', '107')->find_one();
$f113 = Model::factory('FlexModel')->where('modelType', 'F')->where('number', '113')->find_one();
foreach ($ifs as $if) {
$model = StringUtil::endsWith($if->formNoAccent, 'are') ? $f113 : $f107;
// Load an existing lexem only if it has one of the models F113, F107 or T1. Otherwise create a new lexem.
$lexems = Lexem::get_all_by_formNoAccent($if->formNoAccent);
$lexem = null;
foreach ($lexems as $l) {
if ($l->modelType == 'T' || $l->modelType == 'F' && $l->modelNumber == $model->number) {
$lexem = $l;
} else {
if ($this->isLoc && !$l->isLoc) {
FlashMessage::add("Lexemul {$l->formNoAccent} ({$l->modelType}{$l->modelNumber}), care nu este în LOC, nu a fost modificat.", 'info');
}
}
}
if ($lexem) {
$lexem->modelType = 'F';
$lexem->modelNumber = $model->number;
$lexem->restriction = '';
if ($this->isLoc && !$lexem->isLoc) {
$lexem->isLoc = $this->isLoc;
FlashMessage::add("Lexemul {$lexem->formNoAccent}, care nu era în LOC, a fost inclus automat în LOC.", 'info');
}
$lexem->noAccent = false;
$lexem->save();
} else {
$lexem = Lexem::create($if->form, 'F', $model->number, '');
$lexem->isLoc = $this->isLoc;
$lexem->save();
// Also associate the new lexem with the same definitions as $this.
$ldms = LexemDefinitionMap::get_all_by_lexemId($this->id);
foreach ($ldms as $ldm) {
LexemDefinitionMap::associate($lexem->id, $ldm->definitionId);
}
FlashMessage::add("Am creat automat lexemul {$lexem->formNoAccent} (F{$lexem->modelNumber}) și l-am asociat cu toate definițiile verbului.", 'info');
}
$lexem->regenerateParadigm();
}
}
function parseWordField($word, $modelType, $modelNo, $restr)
{
$word = trim($word);
// Look for a slash not included in brackets
$len = mb_strlen($word);
$parCount = 0;
$i = 0;
$found = false;
while ($i < $len && !$found) {
$c = text_getCharAt($word, $i);
if ($c == '[' || $c == '(') {
$parCount++;
} else {
if ($c == ']' || $c == ')') {
$parCount--;
}
}
if ($c == '/' && !$parCount) {
$found = true;
} else {
$i++;
}
}
if ($found) {
$r1 = parseWordField(mb_substr($word, 0, $i), $modelType, $modelNo, $restr);
$r2 = parseWordField(mb_substr($word, $i + 1), $modelType, $modelNo, $restr);
return array_merge($r1, $r2);
}
if (text_endsWith($word, ']')) {
$pos = mb_strrpos($word, '[');
assert($pos !== false);
$extra = mb_substr($word, $pos);
$results = parseWordField(mb_substr($word, 0, $pos), $modelType, $modelNo, $restr);
assert(count($results));
appendExtra($results[count($results) - 1], $extra);
return $results;
}
if (text_endsWith($word, ')')) {
$pos = mb_strrpos($word, '(');
assert($pos !== false);
$extra = mb_substr($word, $pos);
$results = parseWordField(mb_substr($word, 0, $pos), $modelType, $modelNo, $restr);
assert(count($results));
// See if $extra contains a model number. If so, use it on the last model.
list($modelType, $modelNo, $restr) = parseModel($extra);
if ($modelType && $modelNo) {
$results[count($results) - 1]->modelType = $modelType;
$results[count($results) - 1]->modelNumber = $modelNo;
$results[count($results) - 1]->restriction = $restr;
}
appendExtra($results[count($results) - 1], $extra);
// If $extra dictates a part of speech, apply it to all the lexems
if (text_contains($extra, 's.f.inv.') || text_contains($extra, 's.f. în expr.') || text_contains($extra, 's.m.inv.') || text_contains($extra, 's.n.inv.') || text_contains($extra, 'adj.inv.') || text_contains($extra, 'adv.') || text_contains($extra, 'conj.') || text_contains($extra, 'prep.') || text_contains($extra, 'interj.')) {
foreach ($results as $l) {
$l->modelType = 'I';
$l->modelNumber = '1';
$l->restriction = '';
}
}
return $results;
}
$parts = split(',', $word);
if (count($parts) >= 2) {
$results = array();
foreach ($parts as $part) {
$results = array_merge($results, parseWordField($part, $modelType, $modelNo, $restr));
}
return $results;
}
$extra = text_contains($word, '-') ? $word : '';
$word = str_replace('-', '', $word);
$len = mb_strlen($word);
$found = false;
for ($i = 0; $i < $len && !$found; $i++) {
$c = text_getCharAt($word, $i);
if (text_isLowercase($c)) {
$found = true;
$word = text_insert($word, "'", $i);
}
}
$word = text_unicodeToLower($word);
$l = Lexem::create($word, $modelType, $modelNo, $restr);
appendExtra($l, $extra);
$l->isLoc = true;
return array($l);
}
$d->lexicon = text_internalizeWordName($parts[0]);
$d->status = ST_PENDING;
} else {
if (count($parts) == 2 && text_endsWith($parts[1], '/')) {
// Use the first part only, because the second one is the pronunciation
$d->lexicon = text_internalizeWordName($parts[0]);
//print "Using [{$d->lexicon}] for " . mb_substr($d->internalRep, 0, 50) . "\n";
} else {
$d->status = ST_PENDING;
}
}
}
if ($d->lexicon) {
$lexems = Lexem::loadByUnaccented($d->lexicon);
if (!count($lexems)) {
$lexem = Lexem::create($d->lexicon, 'T', '1', '');
$lexem->save();
$lexem->id = db_getLastInsertedId();
$lexem->regenerateParadigm();
$lexems[] = $lexem;
}
$d->save();
$d->id = db_getLastInsertedId();
foreach ($lexems as $l) {
$ldm = LexemDefinitionMap::create($l->id, $d->id);
$ldm->save();
}
} else {
print "Skipping [{$d->internalRep}]\n";
}
}
$l = Lexem::createFromDbRow($dbRow);
$seen++;
$defs = Definition::loadByLexemId($l->id);
if (count($defs)) {
$parts = split(' ', $l->form);
print text_padRight($l->form, 30);
foreach ($parts as $part) {
$part = trim($part);
if (!$part || in_array($part, $stopWords)) {
// Skip common words
continue;
}
print '[';
$baseForms = Lexem::searchWordlists($part, true);
if (!count($baseForms)) {
$baseForm = Lexem::create($part, 'T', '1', '');
$baseForm->comment = "Provine din despărțirea lexemului [{$l->form}]";
$baseForm->noAccent = true;
$baseForm->save();
$baseForm->id = db_getLastInsertedId();
$baseForm->regenerateParadigm();
$baseForms[] = $baseForm;
}
// Associate every definition with every lexem
foreach ($baseForms as $baseForm) {
print $baseForm->form . ' ';
foreach ($defs as $def) {
LexemDefinitionMap::associate($baseForm->id, $def->id);
}
}
print ']';
