Merge pull request 'CLD: Keep the original detected language array' (#1428) from heluecht/friendica-addons:cld into 2023.09-rc

Reviewed-on: https://git.friendi.ca/friendica/friendica-addons/pulls/1428
This commit is contained in:
Hypolite Petovan 2023-10-12 13:36:58 +02:00
commit b11538d195

View file

@ -28,7 +28,7 @@ function cld_detect_languages(array &$data)
$cld2->setPlainText(true); $cld2->setPlainText(true);
$result = $cld2->detect($data['text']); $result = $cld2->detect($data['text']);
if ($data['detected']) { if ($data['detected']) {
$original = array_key_first($data['detected']); $original = array_key_first($data['detected']);
} else { } else {
@ -69,10 +69,9 @@ function cld_detect_languages(array &$data)
if (!$result['is_reliable']) { if (!$result['is_reliable']) {
Logger::debug('Unreliable detection', ['uri-id' => $data['uri-id'], 'original' => $original, 'detected' => $detected, 'name' => $result['language_name'], 'probability' => $result['language_probability'], 'text' => $data['text']]); Logger::debug('Unreliable detection', ['uri-id' => $data['uri-id'], 'original' => $original, 'detected' => $detected, 'name' => $result['language_name'], 'probability' => $result['language_probability'], 'text' => $data['text']]);
return; if (($original == $detected) && ($data['detected'][$original] < $result['language_probability'] / 100)) {
} $data['detected'][$original] = $result['language_probability'] / 100;
}
if ($original == $detected) {
return; return;
} }
@ -83,6 +82,15 @@ function cld_detect_languages(array &$data)
return; return;
} }
Logger::debug('Detected different language', ['uri-id' => $data['uri-id'], 'original' => $original, 'detected' => $detected, 'name' => $result['language_name'], 'probability' => $result['language_probability'], 'text' => $data['text']]); if ($original != $detected) {
$data['detected'] = [$detected => $result['language_probability'] / 100]; Logger::debug('Detected different language', ['uri-id' => $data['uri-id'], 'original' => $original, 'detected' => $detected, 'name' => $result['language_name'], 'probability' => $result['language_probability'], 'text' => $data['text']]);
}
$length = count($data['detected']);
if ($length > 0) {
unset($data['detected'][$detected]);
$data['detected'] = array_merge([$detected => $result['language_probability'] / 100], array_slice($data['detected'], 0, $length - 1));
} else {
$data['detected'] = [$detected => $result['language_probability'] / 100];
}
} }