Skip to content

Commit

Permalink
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Allow multiple aux languages for a given WC call
Browse files Browse the repository at this point in the history 
Rather than restrict to only one additional aux language when calling
WordCheck, allow multiple of them. This is the backend changes for
Task 2062 -- the UI bits need to still be implemented.
@cpeel
cpeel committed Dec 8, 2023
1 parent 965a8d5 commit b087aa7
Showing 3 changed files with 44 additions and 45 deletions.
52 changes: 27 additions & 25 deletions pinc/wordcheck_engine.inc
View file
Original file line number Diff line number Diff line change
@@ -60,16 +60,16 @@ define('WC_PAGE', 4);
* @param string $projectid
* ID of project, needed for project languages and to load the custom
* dictionaries
* @param string $aux_languages
* Auxiliary language to check against.
* @param array $languages
* Languages to check against.
* @param array $adhoc_good_words
* Array of words to treat as good for this invocation only.
*
* @return array
*/
function get_bad_word_freqs_for_project_text($text, $projectid, $aux_language = '', $adhoc_good_words = [])
function get_bad_word_freqs_for_project_text($text, $projectid, $languages, $adhoc_good_words = [])
{
return _get_bad_words_for_project_text('FREQS', $text, $projectid, $aux_language, $adhoc_good_words);
return _get_bad_words_for_project_text('FREQS', $text, $projectid, $languages, $adhoc_good_words);
}

/**
@@ -91,16 +91,16 @@ function get_bad_word_freqs_for_project_text($text, $projectid, $aux_language =
* @param string $projectid
* ID of project, needed for project languages and to load the custom
* dictionaries
* @param string $aux_languages
* Auxiliary language to check against.
* @param array $languages
* Languages to check against.
* @param array $adhoc_good_words
* Array of words to treat as good for this invocation only.
*
* @return array
*/
function get_bad_word_levels_for_project_text($text, $projectid, $aux_language = '', $adhoc_good_words = [])
function get_bad_word_levels_for_project_text($text, $projectid, $languages, $adhoc_good_words = [])
{
return _get_bad_words_for_project_text('LEVELS', $text, $projectid, $aux_language, $adhoc_good_words);
return _get_bad_words_for_project_text('LEVELS', $text, $projectid, $languages, $adhoc_good_words);
}

/**
@@ -127,21 +127,17 @@ function get_bad_word_levels_for_project_text($text, $projectid, $aux_language =
* @param string $projectid
* ID of project, needed for project languages and to load the custom
* dictionaries
* @param string $aux_languages
* Auxiliary language to check against.
* @param array $languages
* Languages to check against.
* @param array $adhoc_good_words
* Array of words to treat as good for this invocation only.
*
* @return array
*/
function _get_bad_words_for_project_text($which_result_values, $text, $projectid, $aux_language, $adhoc_good_words)
function _get_bad_words_for_project_text($which_result_values, $text, $projectid, $languages, $adhoc_good_words)
{
$messages = [];

// Get the list of languages that we'll use. This also ensures that
// $aux_language is a valid language and if not discards it.
$languages = array_values(get_project_languages($projectid, $aux_language));

// And the langcode3s for those languages
[$langcode3s, $new_messages] = get_langcode3s_for_languages($languages);
$messages += $new_messages;
@@ -783,21 +779,21 @@ function load_project_good_word_suggestions($projectid, $start_time = 0)

function load_site_good_words_given_project($projectid)
{
$languages = array_values(get_project_languages($projectid));
$languages = get_project_languages($projectid);
[$langcode3s, $messages] = get_langcode3s_for_languages($languages);
return load_site_good_words($langcode3s);
}

function load_site_bad_words_given_project($projectid)
{
$languages = array_values(get_project_languages($projectid));
$languages = get_project_languages($projectid);
[$langcode3s, $messages] = get_langcode3s_for_languages($languages);
return load_site_bad_words($langcode3s);
}

function load_site_possible_bad_words_given_project($projectid)
{
$languages = array_values(get_project_languages($projectid));
$languages = get_project_languages($projectid);
[$langcode3s, $messages] = get_langcode3s_for_languages($languages);
return load_site_possible_bad_words($langcode3s);
}
@@ -995,9 +991,17 @@ function save_word_list($path, $words)
// -----------------------------------------------------------------------------

/**
* Returns an associative array of all languages for a project
* Returns an array of all languages for a project
*
* @param string $projectid
* ID of project
* @param array $aux_languages
* Any additional languages to append to the list of project languages.
* This function will silently filter out any invalid values.
*
* @return array
*/
function get_project_languages($projectid, $aux_language = null)
function get_project_languages($projectid, $aux_languages = [])
{
$returnArray = [];

@@ -1006,15 +1010,13 @@ function get_project_languages($projectid, $aux_language = null)
} catch (NonexistentProjectException $exception) {
return $returnArray;
}
$languages = $project->languages;
if ($aux_language) {
$languages[] = $aux_language;
}

// filter out unknown languages by removing those without langcode2s
$languages = array_merge($project->languages, $aux_languages);
foreach ($languages as $language) {
$langcode = langcode2_for_langname($language);
if ($langcode) {
$returnArray[$langcode] = $language;
$returnArray[] = $language;
}
}

8 changes: 4 additions & 4 deletions tools/proofers/processtext.php
View file
Original file line number Diff line number Diff line change
@@ -191,7 +191,7 @@
}
// save what we have so far, just in case the spellchecker barfs
$ppage->saveAsInProgress($text_data, $pguser);
$aux_language = '';
$languages = get_project_languages($projectid);
$accepted_words = [];
$text_data = $_POST["text_data"];

@@ -203,7 +203,7 @@
$is_changed = 0;

slim_header(_("WordCheck"), get_wordcheck_page_header_args($user, $ppage));
output_wordcheck_interface($user, $ppage, $text_data, $is_changed, $accepted_words, $aux_language);
output_wordcheck_interface($user, $ppage, $text_data, $is_changed, $accepted_words, $languages);
break;

case 101:
@@ -311,7 +311,7 @@
// User wants to run the page through spellcheck for an another language
// Apply current corrections to text (but don't save the progress)
// and rerun through the spellcheck
$aux_language = $_POST["aux_language"];
$languages = get_project_languages($projectid, [$_POST["aux_language"]]);
$accepted_words = explode(' ', $_POST["accepted_words"]);
$_SESSION["is_header_visible"] = get_integer_param($_POST, 'is_header_visible', 0, 0, 1);
[$text_data, $corrections] = spellcheck_apply_corrections();
@@ -326,7 +326,7 @@
$is_changed = get_integer_param($_POST, 'is_changed', 0, 0, 1);

slim_header(_("WordCheck"), get_wordcheck_page_header_args($user, $ppage));
output_wordcheck_interface($user, $ppage, $text_data, $is_changed, $accepted_words, $aux_language);
output_wordcheck_interface($user, $ppage, $text_data, $is_changed, $accepted_words, $languages);
break;


29 changes: 13 additions & 16 deletions tools/proofers/spellcheck_text.inc
View file
Original file line number Diff line number Diff line change
@@ -24,14 +24,14 @@ use voku\helper\UTF8;
* Id of projected, needed for temp filename and to load the custom dictionaries
* @param text $imagefile
* Image filename, needed for temp filename
* @param text $aux_language
* Auxiliary language to check against
* @param array $languages
* Languages to check against
* @param array $accepted_words
* Array of words that should not be considered misspelled
*
* @return array
*/
function spellcheck_text($orig_text, $projectid, $imagefile, $aux_language, $accepted_words)
function spellcheck_text($orig_text, $projectid, $imagefile, $languages, $accepted_words)
{
global $puncCharacters;

@@ -51,13 +51,8 @@ function spellcheck_text($orig_text, $projectid, $imagefile, $aux_language, $acc
// prepare the text for checking, in this case change all EOLs to spaces
$orig_text = str_replace('[lf]', ' ', $orig_text);

// Handle a UI-ism so that func only needs to handle ''.
if ($aux_language == _("Language")) {
$aux_language = '';
}

[$badWordHash, $languages, $messages] =
get_bad_word_levels_for_project_text($orig_text, $projectid, $aux_language, $accepted_words);
get_bad_word_levels_for_project_text($orig_text, $projectid, $languages, $accepted_words);

[$uncommonScriptWords, $uncommonScripts] =
get_words_with_uncommon_scripts(array_keys(get_distinct_words_in_text($orig_text)));
@@ -393,7 +388,7 @@ $word_check_messages = [
"rerun" => _("Check page against an additional language"),
];

function output_wordcheck_interface($user, $ppage, $text_data, $is_changed, $accepted_words, $aux_language)
function output_wordcheck_interface($user, $ppage, $text_data, $is_changed, $accepted_words, $languages)
{
global $code_url, $word_check_messages;

@@ -421,7 +416,7 @@ function output_wordcheck_interface($user, $ppage, $text_data, $is_changed, $acc

// run the text through the spellcheck - returns the form contents and a list of languages the page was checked against
[$page_contents, $languages, $messages, $highlights] =
spellcheck_text($text_data, $ppage->projectid(), $ppage->imagefile(), $aux_language, $accepted_words);
spellcheck_text($text_data, $ppage->projectid(), $ppage->imagefile(), $languages, $accepted_words);

// start the div containing the link to show/hide the WordCheck header
echo "<div>";
@@ -450,9 +445,10 @@ function output_wordcheck_interface($user, $ppage, $text_data, $is_changed, $acc

// start the first line of the WordCheck header
// printing any dictionaries used
$project_languages = get_project_languages($ppage->projectid());
echo "<p>";
if (count($languages) > 0) {
echo sprintf(_("Dictionaries used: <b>%s</b>."), implode(', ', $languages));
if (count($project_languages) > 0) {
echo sprintf(_("Dictionaries used: <b>%s</b>."), implode(', ', $project_languages));
} else {
echo "<b>" . _("No check against a dictionary has been made.") . "</b>";
}
@@ -471,15 +467,16 @@ function output_wordcheck_interface($user, $ppage, $text_data, $is_changed, $acc
echo "<option value=''>" . _("Language") . "</option>\n";

// get a list of languages with dictionaries installed on the system
$aux_languages = array_diff($languages, $project_languages);
$dict_list = get_languages_with_dictionaries();
asort($dict_list);
foreach ($dict_list as $langcode => $language) {
// skip the languages that we just used.
if (in_array($language, $languages)) {
// skip the project-specific languages that we just used.
if (in_array($language, $project_languages)) {
continue;
}
echo "<option value='" . attr_safe($language) . "'";
if (strcmp($language, $aux_language) == 0) {
if (in_array($language, $aux_languages)) {
echo " selected";
}
echo ">$language</option>\n";

0 comments on commit b087aa7

Please sign in to comment.