function split_token($token_id, $num)
{
//$num is the number of characters (in the beginning) that should become a separate token
check_permission(PERM_ADDER);
if (is_token_covered_by_ne_markup($token_id)) {
throw new Exception("Cannot split token under NE markup");
}
if (!$token_id || !$num) {
throw new UnexpectedValueException();
}
$res = sql_pe("SELECT tf_text, sent_id, pos FROM tokens WHERE tf_id=? LIMIT 1", array($token_id));
if (sizeof($res) == 0) {
throw new Exception();
}
$r = $res[0];
$text1 = trim(mb_substr($r['tf_text'], 0, $num));
$text2 = trim(mb_substr($r['tf_text'], $num));
if (!$text1 || !$text2) {
throw new Exception();
}
sql_begin();
//create revset
$revset_id = create_revset("Token {$token_id} (<" . $r['tf_text'] . ">) split to <{$text1}> and <{$text2}>");
$token_for_form2tf = str_replace('ё', 'е', mb_strtolower($text1));
//update other tokens in the sentence
sql_query("UPDATE tokens SET pos=pos+1 WHERE sent_id = " . $r['sent_id'] . " AND pos > " . $r['pos']);
//create new token and parse
sql_pe("INSERT INTO tokens VALUES(NULL, ?, ?, ?)", array($r['sent_id'], $r['pos'] + 1, $text2));
$parse1 = new MorphParseSet(false, $text1);
$parse2 = new MorphParseSet(false, $text2);
create_tf_revision($revset_id, sql_insert_id(), $parse2->to_xml());
//update old token and parse
sql_pe("DELETE FROM form2tf WHERE tf_id=?", array($token_id));
sql_pe("UPDATE tokens SET tf_text=? WHERE tf_id=? LIMIT 1", array($text1, $token_id));
sql_pe("INSERT INTO form2tf VALUES(?, ?)", array($token_for_form2tf, $token_id));
create_tf_revision($revset_id, $token_id, $parse1->to_xml());
//dropping sentence status
$res = sql_pe("SELECT sent_id FROM tokens WHERE tf_id=? LIMIT 1", array($token_id));
$sent_id = $res[0]['sent_id'];
sql_query("UPDATE sentences SET check_status='0' WHERE sent_id={$sent_id} LIMIT 1");
sql_query("DELETE FROM sentence_check WHERE sent_id={$sent_id}");
sql_commit();
$res = sql_query("SELECT book_id FROM paragraphs WHERE par_id = (SELECT par_id FROM sentences WHERE sent_id={$sent_id} LIMIT 1)");
$r = sql_fetch_array($res);
return array($r['book_id'], $sent_id);
}
function update_pending_token($token_id, $rev_id, $revset_id = 0, $smart = false)
{
check_permission(PERM_DICT);
if (!check_safe_token_update($token_id, $rev_id)) {
throw new Exception("Update forbidden");
}
// ok, now we can safely update
$res = sql_pe("SELECT tf_text FROM tokens WHERE tf_id=? LIMIT 1", array($token_id));
$token_text = $res[0]['tf_text'];
$res = sql_pe("SELECT rev_text FROM tf_revisions WHERE tf_id=? AND is_last=1 LIMIT 1", array($token_id));
$previous_rev = $res[0]['rev_text'];
if ($smart) {
$parse = new MorphParseSet($previous_rev);
smart_update_pending_token($parse, $rev_id);
} else {
$parse = new MorphParseSet(false, $token_text);
}
$new_rev = $parse->to_xml();
// do nothing if nothing changed
if ($previous_rev == $new_rev) {
forget_pending_token($token_id, $rev_id);
return true;
}
sql_begin();
if (!$revset_id) {
$revset_id = create_revset("Update tokens from dictionary");
}
create_tf_revision($revset_id, $token_id, $new_rev);
forget_pending_token($token_id, $rev_id);
delete_samples_by_token_id($token_id);
sql_commit();
}
function sentence_save($sent_id)
{
check_permission(PERM_DISAMB);
if (!$sent_id) {
throw new UnexpectedValueException();
}
$flag = $_POST['var_flag'];
//what morphovariants are checked as possible (array of arrays)
$dict = $_POST['dict_flag'];
//whether this token has been reloaded from the dictionary (array)
$res = sql_query("\n SELECT tf_id, tf_text, rev_text\n FROM tokens\n LEFT JOIN tf_revisions\n USING (tf_id)\n WHERE sent_id={$sent_id}\n AND is_last = 1\n ORDER BY `pos`\n ");
while ($r = sql_fetch_array($res)) {
$tokens[$r['tf_id']] = array($r['tf_text'], $r['rev_text']);
}
$matches = array();
$all_changes = array();
if (count($flag) != count($tokens)) {
throw new Exception();
}
sql_begin();
foreach ($tokens as $tf_id => $v) {
list($tf_text, $base_xml) = $v;
//substitute the last revision's xml for one from dictionary if relevant
if (isset($dict[$tf_id]) && $dict[$tf_id] == 1) {
$parse = new MorphParseSet(false, $tf_text, false, true);
} else {
$parse = new MorphParseSet($base_xml);
}
if (sizeof($parse->parses) == 0) {
throw new Exception();
}
// flags quantity check
if (sizeof($parse->parses) != sizeof($flag[$tf_id])) {
throw new Exception();
}
// XXX this is bad since order of parse selection from db
// is not guaranteed to be consistent
$parse->filter_by_parse_index(prepare_parse_indices($flag[$tf_id]));
$new_xml = $parse->to_xml();
if ($base_xml != $new_xml) {
//something's changed
array_push($all_changes, array($tf_id, $new_xml));
}
}
if (count($all_changes) > 0) {
$revset_id = create_revset($_POST['comment']);
foreach ($all_changes as $v) {
create_tf_revision($revset_id, $v[0], $v[1]);
}
}
sql_query("UPDATE sentences SET check_status='1' WHERE sent_id={$sent_id} LIMIT 1");
sql_commit();
}
function addtext_add($text, $sentences, $book_id, $par_num)
{
check_permission(PERM_ADDER);
if (!$text || !$book_id || !$par_num) {
throw new UnexpectedValueException();
}
if (sizeof(sql_pe("SELECT book_id FROM books WHERE parent_id=?", array($book_id))) > 0) {
throw new UnexpectedValueException("Can't add paragraphs to a text having subtexts");
}
sql_begin();
$revset_id = create_revset();
$sent_count = 0;
$pars = split2paragraphs($text);
// move the following paragraphs
sql_query("UPDATE paragraphs SET pos=pos+" . sizeof($pars) . " WHERE book_id = {$book_id} AND pos >= {$par_num}");
$par_ins = sql_prepare("INSERT INTO `paragraphs` VALUES(NULL, ?, ?)");
$sent_ins = sql_prepare("INSERT INTO `sentences` VALUES(NULL, ?, ?, ?, 0)");
$token_ins = sql_prepare("INSERT INTO `tokens` VALUES(NULL, ?, ?, ?)");
foreach ($pars as $par) {
//adding a paragraph
sql_execute($par_ins, array($book_id, $par_num++));
$par_id = sql_insert_id();
$sent_num = 1;
$sents = split2sentences($par);
foreach ($sents as $sent) {
if (!preg_match('/\\S/', $sent)) {
continue;
}
//adding a sentence
sql_execute($sent_ins, array($par_id, $sent_num++, trim($sent)));
$sent_id = sql_insert_id();
sql_query("INSERT INTO sentence_authors VALUES({$sent_id}, " . $_SESSION['user_id'] . ", " . time() . ")");
$token_num = 1;
$tokens = explode('^^', $sentences[$sent_count++]);
foreach ($tokens as $token) {
if (trim($token) === '') {
continue;
}
//adding a textform
sql_execute($token_ins, array($sent_id, $token_num++, trim($token)));
$tf_id = sql_insert_id();
//adding a revision
$parse = new MorphParseSet(false, trim($token));
create_tf_revision($revset_id, $tf_id, $parse->to_xml());
}
}
}
sql_commit();
}