function _get_word_list($projectid, $timeCutoff)
{
$messages = array();
// load the suggestions
$suggestions = load_project_good_word_suggestions($projectid, $timeCutoff);
if (!is_array($suggestions)) {
$messages[] = sprintf(_("Unable to load suggestions: %s"), $suggestions);
return array(array(), array(), array(), array(), array(), array(), $messages);
}
if (count($suggestions) == 0) {
return array(array(), array(), array(), array(), array(), array(), $messages);
}
// load project good words
$project_good_words = load_project_good_words($projectid);
// load project bad words
$project_bad_words = load_project_bad_words($projectid);
// get the latest project text of all pages up to last possible round
$last_possible_round = get_Round_for_round_number(MAX_NUM_PAGE_EDITING_ROUNDS);
$pages_res = page_info_query($projectid, $last_possible_round->id, 'LE');
$all_words_w_freq = get_distinct_words_in_text(get_page_texts($pages_res));
// array to hold all words
$all_suggestions = array();
$round_page_count = array();
// parse the suggestions complex array
// it is in the format: $suggestions[$round][$pagenum]=$wordsArray
foreach ($suggestions as $round => $pageArray) {
$round_suggestions = array();
foreach ($pageArray as $page => $words) {
// add the words to the per-round array
$round_suggestions = array_merge($round_suggestions, $words);
// add the words to the combined array too
$all_suggestions = array_merge($all_suggestions, $words);
@$round_page_count[$round]++;
}
// remove any words already on the project's good or bad words lists
$round_suggestions = array_diff($round_suggestions, array_merge($project_good_words, $project_bad_words));
// get the suggestion occurrences
$round_suggestions_w_occurrences[$round] = generate_frequencies($round_suggestions);
// get suggestion with project word frequency
$round_suggestions_w_freq[$round] = array_intersect_key($all_words_w_freq, array_flip($round_suggestions));
// multisort screws up all-numeric words so we need to preprocess first
prep_numeric_keys_for_multisort($round_suggestions_w_freq[$round]);
// sort the list by frequency, then by word
array_multisort(array_values($round_suggestions_w_freq[$round]), SORT_DESC, array_map('strtolower', array_keys($round_suggestions_w_freq[$round])), SORT_ASC, $round_suggestions_w_freq[$round]);
}
// now, remove any words that are already on the project's good or bad words lists
$all_suggestions = array_diff($all_suggestions, array_merge($project_good_words, $project_bad_words));
// get the number of suggestion occurrences
$all_suggestions_w_occurrences = generate_frequencies($all_suggestions);
// $all_suggestions doesn't have frequency info,
// so start with the info in $all_words_w_freq,
// and extract the items where the key matches a key in $all_suggestions.
$all_suggestions_w_freq = array_intersect_key($all_words_w_freq, array_flip($all_suggestions));
// multisort screws up all-numeric words so we need to preprocess first
prep_numeric_keys_for_multisort($all_suggestions_w_freq);
// sort the list by frequency, then by word
array_multisort(array_values($all_suggestions_w_freq), SORT_DESC, array_map('strtolower', array_keys($all_suggestions_w_freq)), SORT_ASC, $all_suggestions_w_freq);
// get a list of all rounds
$rounds = array_keys($round_suggestions_w_freq);
return array($all_suggestions_w_freq, $all_suggestions_w_occurrences, $round_suggestions_w_freq, $round_suggestions_w_occurrences, $rounds, $round_page_count, $messages);
}
function _get_word_list($projectid)
{
$messages = array();
// get the latest project text of all pages up to last possible round
$last_possible_round = get_Round_for_round_number(MAX_NUM_PAGE_EDITING_ROUNDS);
$pages_res = page_info_query($projectid, $last_possible_round->id, 'LE');
$all_words_w_freq = get_distinct_words_in_text(get_page_texts($pages_res));
// load site word lists for project languages
$site_possible_bad_words = load_site_possible_bad_words_given_project($projectid);
// now, remove any words that are already on the project's bad word list
$site_possible_bad_words = array_diff($site_possible_bad_words, load_project_bad_words($projectid));
// $site_possible_bad_words doesn't have frequency info,
// so start with the info in $all_words_w_freq,
// and extract the items where the key matches a key in $bad_words.
$bad_words_w_freq = array_intersect_key($all_words_w_freq, array_flip($site_possible_bad_words));
// multisort screws up all-numeric words so we need to preprocess first
prep_numeric_keys_for_multisort($bad_words_w_freq);
// sort the list by frequency, then by word
array_multisort(array_values($bad_words_w_freq), SORT_DESC, array_map('strtolower', array_keys($bad_words_w_freq)), SORT_ASC, $bad_words_w_freq);
return array($bad_words_w_freq, $messages);
}
function _get_word_list($projectid)
{
global $aspell_temp_dir;
$ocr_filename = "{$aspell_temp_dir}/{$projectid}_ocr.txt";
$latest_filename = "{$aspell_temp_dir}/{$projectid}_latest.txt";
$messages = array();
// get the OCR text
// Note: If the code is changed to allow selecting of arbitrary round text
// instead of just the OCR round, edit_project_word_lists.php should
// be updated to allow this page to be accessed for those type-in
// projects with no OCR text.
$pages_res = page_info_query($projectid, '[OCR]', 'LE');
$all_page_text = get_page_texts($pages_res);
// remove any formatting tags and add a final \r\n to each page-text
// to ensure that there is whitespace between pages so they don't run together
$all_page_text = preg_replace(array('#<[/]?\\w+>#', '#$#'), array('', "\r\n"), $all_page_text);
file_put_contents($ocr_filename, $all_page_text);
// get the latest project text of all pages up to last possible round
$last_possible_round = get_Round_for_round_number(MAX_NUM_PAGE_EDITING_ROUNDS);
$pages_res = page_info_query($projectid, $last_possible_round->id, 'LE');
$all_page_text = get_page_texts($pages_res);
// remove any formatting tags and add a final \r\n to each page-text
// to ensure that there is whitespace between pages so they don't run together
$all_page_text = preg_replace(array('#<[/]?\\w+>#', '#$#'), array('', "\r\n"), $all_page_text);
file_put_contents($latest_filename, $all_page_text);
$all_words_w_freq = get_distinct_words_in_text($all_page_text);
// clean up unused variables
unset($all_page_text);
// make external call to wdiff
exec("wdiff -3 {$ocr_filename} {$latest_filename}", $wdiff_output, $return_code);
// check to see if wdiff wasn't found to execute
if ($return_code == 127) {
die("Error invoking wdiff to do the diff analysis. Perhaps it is not installed.");
}
if ($return_code == 2) {
die("Error reported from wdiff while attempting to do the diff analysis.");
}
// clean up the temporary files
if (is_file($ocr_filename)) {
unlink($ocr_filename);
}
if (is_file($latest_filename)) {
unlink($latest_filename);
}
// specify the separator between the wdiff segments
$separator = '======================================================================';
$possible_scannos_w_correction = array();
$possible_scannos_w_count = array();
// parse the incoming data one segment at a time
// from the original datastream to conserve memory
$lineIndex = 0;
$totalLines = count($wdiff_output);
while ($lineIndex < $totalLines) {
// pull the next segment
$segment = "";
while ($lineIndex <= $totalLines) {
$line = $wdiff_output[$lineIndex];
$lineIndex++;
if ($line == $separator) {
break;
}
$segment .= "{$line}\n";
}
// note that we're handling the case where two adjacent
// words are updated
$ocr_words = $latest_words = array();
// pull out the original word(s)
if (preg_match("/\\[-(.*?)-\\]/", $segment, $matches)) {
$ocr_words = $matches[1];
$ocr_words = get_all_words_in_text($ocr_words);
}
// if we don't have any ocr_words (probably because
// the correction spanned lines) then don't bother
// continuing with this segment
if (!count($ocr_words)) {
continue;
}
// pull out the replacement(s)
if (preg_match("/{\\+(.*?)\\+}/", $segment, $matches)) {
$latest_words = $matches[1];
$latest_words = get_all_words_in_text($latest_words);
}
// if the number of words isn't the same between the two
// bail since we don't handle that case yet
if (count($ocr_words) != count($latest_words)) {
continue;
}
// process the words, handles multi-words strings
for ($index = 0; $index < count($ocr_words); $index++) {
$ocr_word = $ocr_words[$index];
$latest_word = $latest_words[$index];
// if the words are the same or one of them empty, skip it
if ($ocr_word == $latest_word || empty($ocr_word) || empty($latest_word)) {
continue;
}
$possible_scannos_w_correction[$ocr_word] = $latest_word;
@$possible_scannos_w_count[$ocr_word]++;
}
}
// $wdiff_output can be very large
// so unset it here to be nice for the rest of the function
unset($wdiff_output);
$possible_scannos = array_keys($possible_scannos_w_correction);
// create a string of words to run through WordCheck
$text_to_check = implode(" ", $possible_scannos);
// run the list through WordCheck to see which it would flag
list($possible_scannos_via_wordcheck, $languages, $messages) = get_bad_words_for_text($text_to_check, $projectid, 'all', '', array(), 'FREQS');
// load site words
$site_bad_words = load_site_bad_words_given_project($projectid);
// load the project bad words
$project_bad_words = load_project_bad_words($projectid);
// remove words that WordCheck would flag
$possible_scannos = array_diff($possible_scannos, array_keys($possible_scannos_via_wordcheck));
// remove any scannos already on the site and project bad word lists
$possible_scannos = array_diff($possible_scannos, $site_bad_words, $project_bad_words);
// $possible_scannos doesn't have frequency info,
// so start with the info in $all_words_w_freq,
// and extract the items where the key matches a key in $possible_scannos
$possible_scannos_w_freq = array_intersect_key($all_words_w_freq, array_flip($possible_scannos));
$percent_changed = array();
foreach ($possible_scannos as $word) {
$count = $possible_scannos_w_count[$word];
$totalInstances = @$possible_scannos_w_freq[$word] + $count;
$percent_changed[$word] = sprintf("%0.2f", $count / $totalInstances * 100);
if ($percent_changed[$word] >= 100 && $totalInstances == 1) {
unset($percent_changed[$word]);
}
}
// multisort screws up all-numeric words so we need to preprocess first
prep_numeric_keys_for_multisort($percent_changed);
// sort the list by frequency, then by word
array_multisort(array_values($percent_changed), SORT_DESC, array_map('strtolower', array_keys($percent_changed)), SORT_ASC, $percent_changed);
return array($percent_changed, $possible_scannos_w_freq, $messages, $possible_scannos_w_correction, $possible_scannos_w_count);
}
function save_to_db()
{
global $projects_dir, $pguser;
$postednum_str = $this->postednum == "" ? "NULL" : "'{$this->postednum}'";
// Call addslashes() on any members of $this that might contain
// single-quotes/apostrophes (because they are unescaped, and
// would otherwise break the query).
$common_project_settings = "\n t_last_edit = UNIX_TIMESTAMP(),\n up_projectid = '{$this->up_projectid}',\n nameofwork = '" . addslashes($this->nameofwork) . "',\n authorsname = '" . addslashes($this->authorsname) . "',\n language = '{$this->language}',\n genre = '{$this->genre}',\n difficulty = '{$this->difficulty_level}',\n special_code = '{$this->special_code}',\n clearance = '" . addslashes($this->clearance) . "',\n comments = '" . addslashes($this->comments) . "',\n image_source = '{$this->image_source}',\n scannercredit = '" . addslashes($this->scannercredit) . "',\n checkedoutby = '{$this->checkedoutby}',\n postednum = {$postednum_str},\n image_preparer = '{$this->image_preparer}',\n text_preparer = '{$this->text_preparer}',\n extra_credits = '" . addslashes($this->extra_credits) . "',\n deletion_reason= '" . addslashes($this->deletion_reason) . "'\n ";
$pm_setter = '';
if (user_is_a_sitemanager()) {
// can change PM
$pm_setter = " username = '******',";
} else {
if (isset($this->clone_projectid)) {
// cloning a project. The PM should be the same as
// that of the project being cloned, if the user
// isn't an SA
$res = mysql_query("\n SELECT username\n FROM projects\n WHERE projectid='{$this->clone_projectid}'\n ") or die(mysql_error());
list($projectmanager) = mysql_fetch_row($res);
$pm_setter = " username = '******',";
}
}
if (isset($this->projectid)) {
// We are updating an already-existing project.
// needn't change $pm_setter, as there is no change if the user
// isn't an SA
// find out what we are changing from
$old_pih = new ProjectInfoHolder();
$fatal_error = $old_pih->set_from_db(TRUE, $this->projectid);
if ($fatal_error != '') {
$fatal_error = _('site error') . ': ' . $fatal_error;
echo "<br><center><font size='+1' color='#ff0000'><b>{$fatal_error}</b></font></center>";
exit;
}
$changed_fields = get_changed_fields($this, $old_pih);
// We're particularly interested in knowing
// when the project comments change.
if (!in_array('comments', $changed_fields)) {
// no change
$tlcc_setter = '';
} else {
// changed!
$tlcc_setter = 't_last_change_comments = UNIX_TIMESTAMP(),';
}
// We also want to know if the edit is resulting in the project
// effectively being checked out to a new PPer
if ($old_pih->state == PROJ_POST_FIRST_CHECKED_OUT && in_array('checkedoutby', $changed_fields)) {
$md_setter = 'modifieddate = UNIX_TIMESTAMP(),';
$PPer_checkout = TRUE;
} else {
$md_setter = '';
$PPer_checkout = FALSE;
}
// Update the projects database with the updated info
mysql_query("\n UPDATE projects SET\n {$pm_setter}\n {$tlcc_setter}\n {$md_setter}\n {$common_project_settings}\n WHERE projectid='{$this->projectid}'\n ") or die(mysql_error());
$details1 = implode(' ', $changed_fields);
if ($details1 == '') {
// There are no changed fields.
// Don't just save '' for the details1 column,
// because then do_history() won't be able to distinguish
// this case (no changed fields) from old cases
// (edit occurred before we started recording changed fields).
// Instead, use a special value.
$details1 = 'NONE';
}
$e = log_project_event($this->projectid, $GLOBALS['pguser'], 'edit', $details1);
if (!empty($e)) {
die($e);
}
if ($PPer_checkout) {
// we fake the project transition...
$e = log_project_event($this->projectid, $GLOBALS['pguser'], 'transition', PROJ_POST_FIRST_CHECKED_OUT, PROJ_POST_FIRST_CHECKED_OUT, $this->checkedoutby);
if (!empty($e)) {
die($e);
}
}
// Update the MARC record with any info we've received.
$project = new Project($this->projectid);
$marc_record = $project->load_marc_record();
$this->update_marc_record_from_post($marc_record);
$project->save_marc_record($marc_record);
} else {
// We are creating a new project
$this->projectid = uniqid("projectID");
// The project ID
if ('' == $pm_setter) {
$pm_setter = "username = '******',";
}
// Insert a new row into the projects table
mysql_query("\n INSERT INTO projects\n SET\n projectid = '{$this->projectid}',\n {$pm_setter}\n state = '" . PROJ_NEW . "',\n modifieddate = UNIX_TIMESTAMP(),\n t_last_change_comments = UNIX_TIMESTAMP(),\n {$common_project_settings}\n ") or die(mysql_error());
$e = log_project_event($this->projectid, $GLOBALS['pguser'], 'creation');
if (!empty($e)) {
die($e);
}
$e = project_allow_pages($this->projectid);
if (!empty($e)) {
die($e);
}
// Make a directory in the projects_dir for this project
mkdir("{$projects_dir}/{$this->projectid}", 0777) or die("System error: unable to mkdir '{$projects_dir}/{$this->projectid}'");
chmod("{$projects_dir}/{$this->projectid}", 0777);
// Do MARC record manipulations
$project = new Project($this->projectid);
$marc_record = new MARCRecord();
// Save original MARC record, if provided
$yaz_array = unserialize(base64_decode($this->original_marc_array_encd));
if ($yaz_array !== FALSE) {
$marc_record->load_yaz_array($yaz_array);
$project->init_marc_record($marc_record);
// Update the MARC record with data from POST
$this->update_marc_record_from_post($marc_record);
$project->save_marc_record($marc_record);
}
// Create the project's 'good word list' and 'bad word list'.
if (isset($this->clone_projectid)) {
// We're creating a project via cloning.
// Copy the original project's word-lists.
$good_words = load_project_good_words($this->clone_projectid);
if (is_string($good_words)) {
// It's an error message.
echo "{$good_words}<br>\n";
$good_words = array();
}
$bad_words = load_project_bad_words($this->clone_projectid);
if (is_string($bad_words)) {
// It's an error message.
echo "{$bad_words}<br>\n";
$bad_words = array();
}
} else {
// We're creating a project by means other than cloning
// (from_nothing, from_marc_record, from_uberproject).
// Initialize its GWL and BWL to empty.
$good_words = array();
$bad_words = array();
}
save_project_good_words($this->projectid, $good_words);
save_project_bad_words($this->projectid, $bad_words);
}
// Create/update the Dublin Core file for the project.
// When we get here, the project's database entry has been fully
// updated, so we can create a Project object and allow it
// to pull the relevant fields from the database.
$project = new Project($this->projectid);
$project->create_dc_xml_oai($marc_record);
// If the project has been posted to PG, make the appropriate transition.
if ($this->posted) {
$err = project_transition($this->projectid, PROJ_SUBMIT_PG_POSTED, $pguser);
if ($err != '') {
echo "{$err}<br>\n";
exit;
}
}
}
$cutoffOptions = array(1, 2, 3, 4, 5, 10, 25, 50);
// what is the initial cutoff frequency?
$initialFreq = getInitialCutoff($freqCutoff, $cutoffOptions, $bad_words_w_freq);
// echo page support text, like JS and stylesheets
echo_cutoff_script($cutoffOptions, $instances);
echo_word_freq_style();
echo "<p>{$page_text}</p>";
echo_page_instruction_text("good", $format);
echo_download_text($projectid, $format);
echo_any_warnings_errors($messages);
echo_cutoff_text($initialFreq, $cutoffOptions);
$context_array = build_context_array_links($bad_words_w_freq, $projectid);
// load site bad words
$site_bad_words = load_site_bad_words_given_project($projectid);
// load project bad words
$project_bad_words = load_project_bad_words($projectid);
$word_notes = array();
foreach ($site_bad_words as $word) {
$word_notes[$word] = _("On site BWL");
}
foreach ($project_bad_words as $word) {
$word_notes[$word] = _("On project BWL");
}
$context_array["[[TITLE]]"] = _("Show Context");
$word_notes["[[TITLE]]"] = _("Notes");
if (isset($update_status) && $update_status !== "Success") {
echo_any_warnings_errors(array($update_status));
}
$word_checkbox = build_checkbox_array($bad_words_w_freq);
echo_checkbox_selects(count($bad_words_w_freq));
$checkbox_form["projectid"] = $projectid;
function _handle_action($action, $list_type, $language, $cutoff, $lang_match)
{
$display_list = FALSE;
switch ($action) {
case "show":
$word_freq = array();
$total_projects = 0;
$total_projects_with_words = 0;
// figure out what kind of language matching we're going to use
$where_clause = "";
switch ($lang_match) {
case "exact":
$where_clause = "language = '{$language}'";
break;
case "primary":
$where_clause = "language like '{$language}%'";
break;
case "any":
$where_clause = "language like '%{$language}%'";
break;
default:
die("Unknown language match used: {$lang_match}");
}
// loop through all projects that use $language
$res = mysql_query("\n SELECT projectid\n FROM projects\n WHERE {$where_clause}\n ");
while (list($projectid) = mysql_fetch_row($res)) {
if ($list_type == "good") {
$words = load_project_good_words($projectid);
} elseif ($list_type == "bad") {
$words = load_project_bad_words($projectid);
} else {
die("Unknown list type: {$list_type}");
}
foreach ($words as $word) {
@$word_freq[$word]++;
}
if (count($words)) {
$total_projects_with_words++;
}
$total_projects++;
}
mysql_free_result($res);
// sort the results
arsort($word_freq);
// show the results
echo "<pre>";
echo _("Language") . ": {$language}<br>";
echo sprintf(_("Language match type: %s"), $lang_match) . "<br>";
echo sprintf(_("Word list type: %s"), $list_type) . "<br>";
echo sprintf(_("Cutoff percentage: %d%%"), $cutoff) . "<br>";
echo sprintf(_("Total projects matching language: %d"), $total_projects) . "<br>";
echo sprintf(_("Total projects with word lists: %d"), $total_projects_with_words) . "<br>";
echo "<br>";
echo _("Note: Percentages are calculated as frequency over the total number of projects with word lists.") . "<br>";
echo "<br>";
echo sprintf("%20s %5s %s<br>", _("Word"), _("Count"), _("Frequency"));
foreach ($word_freq as $word => $freq) {
$percentage = $freq / $total_projects_with_words * 100;
if ($percentage < $cutoff) {
break;
}
echo sprintf("%20s %5d (%-3.2f%%)<br>", $word, $freq, $percentage);
}
echo "</pre>";
break;
case "list":
$display_list = TRUE;
break;
default:
die("Invalid action encountered.");
}
return $display_list;
}
function merge_wordcheck_files($from_id, $to_id)
{
global $projects_dir;
// good words
$from_words = load_project_good_words($from_id);
$to_words = load_project_good_words($to_id);
$to_words = array_merge($to_words, $from_words);
save_project_good_words($to_id, $to_words);
// crying out for some abstraction here?
// bad words
$from_words = load_project_bad_words($from_id);
$to_words = load_project_bad_words($to_id);
$to_words = array_merge($to_words, $from_words);
save_project_bad_words($to_id, $to_words);
// suggestions
// the file format is complicated and may change
// so we take the sledgehammer approach, as suggested by cpeel...
$from_path = "{$projects_dir}/{$from_id}/good_word_suggestions.txt";
if (!is_file($from_path)) {
// The file does not exist.
// Treat that the same as if it existed and was empty.
$from_suggs = "";
} else {
$from_suggs = file_get_contents($from_path);
}
$to_path = "{$projects_dir}/{$to_id}/good_word_suggestions.txt";
if (!is_file($to_path)) {
// The file does not exist.
// Treat that the same as if it existed and was empty.
$to_suggs = "";
} else {
$to_suggs = file_get_contents($to_path);
}
file_put_contents($to_path, $to_suggs . $from_suggs);
// we're assuming the projects are in unavailable or waiting, so there
// is going to be no need to put locks on the files or anything fancy
}
function set_from_files($load_good_words = true, $load_bad_words = true)
{
$errors = array();
if ($load_good_words) {
$gwl_object = get_project_word_file($this->projectid, "good");
$this->gwl_timestamp = $gwl_object->mod_time;
$good_words = load_project_good_words($this->projectid);
if (is_string($good_words)) {
array_push($errors, $good_words);
$this->good_words = '';
} else {
$this->good_words = implode("\n", $good_words);
}
}
if ($load_bad_words) {
$bwl_object = get_project_word_file($this->projectid, "bad");
$this->bwl_timestamp = $bwl_object->mod_time;
$bad_words = load_project_bad_words($this->projectid);
if (is_string($bad_words)) {
array_push($errors, $bad_words);
$this->bad_words = '';
} else {
$this->bad_words = implode("\n", $bad_words);
}
}
return $errors;
}
