Neues Modul: Qualitätsmanagement
All checks were successful
CI / PHP 8.3 / mariadb:10.6 (pull_request) Successful in 2m46s
CI / PHP 8.4 / mariadb:10.6 (pull_request) Successful in 2m51s
CI / PHP 8.3 / mariadb:11.4 (pull_request) Successful in 3m2s
CI / PHP 8.4 / mariadb:11.4 (pull_request) Successful in 2m55s
CI / PHP 8.3 / mysql:8.0 (pull_request) Successful in 4m15s
CI / PHP 8.4 / mysql:8.0 (pull_request) Successful in 4m16s
CI / PHP 8.4 / mysql:8.4 (pull_request) Successful in 3m44s
CI / PHP 8.3 / mysql:8.4 (pull_request) Successful in 3m46s
CI / Build-Artefakte aktuell (pull_request) Successful in 52s
All checks were successful
CI / PHP 8.3 / mariadb:10.6 (pull_request) Successful in 2m46s
CI / PHP 8.4 / mariadb:10.6 (pull_request) Successful in 2m51s
CI / PHP 8.3 / mariadb:11.4 (pull_request) Successful in 3m2s
CI / PHP 8.4 / mariadb:11.4 (pull_request) Successful in 2m55s
CI / PHP 8.3 / mysql:8.0 (pull_request) Successful in 4m15s
CI / PHP 8.4 / mysql:8.0 (pull_request) Successful in 4m16s
CI / PHP 8.4 / mysql:8.4 (pull_request) Successful in 3m44s
CI / PHP 8.3 / mysql:8.4 (pull_request) Successful in 3m46s
CI / Build-Artefakte aktuell (pull_request) Successful in 52s
- Kategorien mit farbigen Ordnerkacheln; Farbe und Symbol je Ordner (Auswahl oder eigener Upload, SVG bereinigt), an Unterordner vererbt - Ordnerseite: Einführungstext, Ansprechpartner (Profilfelder wählbar), Unterordner, Liste mit Dokument, Version, Stand, Größe, Aktion - Einträge: Datei (Lesefassung + bearbeitbare Vorlage), Link, Online-Formular; neue Fassungen archivieren die bisherigen - PDF-Betrachter im Portal (pdf.js) - Eigene Volltextsuche (PDF, Word, Excel, PowerPoint, OpenDocument) mit Textausschnitt; optionale Texterkennung per Tesseract - Rollen-Ausschluss für das ganze Modul, Pflegerecht qm.manage, Zielgruppen je Kategorie - Vorbereitet: Wiedervorlage, Verantwortliche, Lesebestätigungen - Abhängigkeit smalot/pdfparser Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This commit is contained in:
parent
f78cc96be7
commit
b3a057e27c
218 changed files with 13478 additions and 358 deletions
163
plugins-core/qm/src/Text/Ocr.php
Normal file
163
plugins-core/qm/src/Text/Ocr.php
Normal file
|
|
@ -0,0 +1,163 @@
|
|||
<?php
|
||||
|
||||
namespace Inhaus\Plugins\Qm\Text;
|
||||
|
||||
use App\Support\Settings;
|
||||
use Imagick;
|
||||
use Throwable;
|
||||
|
||||
/**
|
||||
* Texterkennung (OCR) für eingescannte PDFs – optional, wenn der Server sie kann:
|
||||
* Tesseract (Programm „tesseract“ mit deutschen Sprachdaten) und ein Umwandler für
|
||||
* PDF-Seiten in Bilder (Programm „pdftoppm“ aus poppler-utils oder PHP-Imagick mit
|
||||
* Ghostscript). Auf Shared Hosting meist nicht vorhanden – dann bleiben solche
|
||||
* Dateien in der Suche nur über ihren Titel auffindbar.
|
||||
*
|
||||
* Einrichtung (Debian/Ubuntu): apt install tesseract-ocr tesseract-ocr-deu poppler-utils
|
||||
*/
|
||||
class Ocr
|
||||
{
|
||||
public const SETTING_ENABLED = 'qm.ocr.enabled';
|
||||
|
||||
public const SETTING_BINARY = 'qm.ocr.tesseract';
|
||||
|
||||
public const MAX_PAGES = 40;
|
||||
|
||||
public function __construct(private readonly Settings $settings) {}
|
||||
|
||||
public function enabled(): bool
|
||||
{
|
||||
return (bool) $this->settings->get(self::SETTING_ENABLED, false);
|
||||
}
|
||||
|
||||
/**
|
||||
* Was auf diesem Server vorhanden ist.
|
||||
*
|
||||
* @return array{exec: bool, // Programme dürfen aufgerufen werden (proc_open)
|
||||
* tesseract: ?string, languages: list<string>, rasterizer: ?string}
|
||||
*/
|
||||
public function status(): array
|
||||
{
|
||||
$exec = function_exists('proc_open') && ! in_array('proc_open', array_map('trim', explode(',', (string) ini_get('disable_functions'))), true);
|
||||
$tesseract = null;
|
||||
$languages = [];
|
||||
|
||||
if ($exec) {
|
||||
$binary = $this->binary();
|
||||
$version = $this->run([$binary, '--version']);
|
||||
|
||||
if ($version !== null && str_contains(strtolower($version), 'tesseract')) {
|
||||
$tesseract = trim(strtok($version, "\n"));
|
||||
$languages = array_values(array_filter(array_map('trim', array_slice(explode("\n", (string) $this->run([$binary, '--list-langs'])), 1))));
|
||||
}
|
||||
}
|
||||
|
||||
return ['exec' => $exec, 'tesseract' => $tesseract, 'languages' => $languages, 'rasterizer' => $this->rasterizer($exec)];
|
||||
}
|
||||
|
||||
public function available(): bool
|
||||
{
|
||||
$status = $this->status();
|
||||
|
||||
return $status['tesseract'] !== null && $status['rasterizer'] !== null;
|
||||
}
|
||||
|
||||
/**
|
||||
* Text einer PDF-Datei per Texterkennung.
|
||||
*/
|
||||
public function recognize(string $pdf): string
|
||||
{
|
||||
$dir = sys_get_temp_dir().'/inhaus-ocr-'.bin2hex(random_bytes(6));
|
||||
@mkdir($dir, 0700, true);
|
||||
|
||||
try {
|
||||
$images = $this->rasterize($pdf, $dir);
|
||||
$languages = in_array('deu', $this->status()['languages'], true) ? 'deu+eng' : 'eng';
|
||||
$text = '';
|
||||
|
||||
foreach ($images as $image) {
|
||||
$text .= $this->run([$this->binary(), $image, 'stdout', '-l', $languages, '--psm', '1'])."\n";
|
||||
}
|
||||
|
||||
return TextExtractor::normalize($text);
|
||||
} finally {
|
||||
array_map('unlink', glob($dir.'/*') ?: []);
|
||||
@rmdir($dir);
|
||||
}
|
||||
}
|
||||
|
||||
private function binary(): string
|
||||
{
|
||||
$binary = trim((string) $this->settings->get(self::SETTING_BINARY, '')) ?: 'tesseract';
|
||||
|
||||
// Nur ein Programmname oder ein absoluter Pfad – keine Parameter.
|
||||
return preg_match('#^(/[\w./-]+|[\w.-]+)$#', $binary) ? $binary : 'tesseract';
|
||||
}
|
||||
|
||||
private function rasterizer(bool $exec): ?string
|
||||
{
|
||||
if ($exec && $this->run(['pdftoppm', '-v']) !== null) {
|
||||
return 'pdftoppm';
|
||||
}
|
||||
|
||||
try {
|
||||
if (class_exists(Imagick::class) && in_array('PDF', Imagick::queryFormats('PDF'), true)) {
|
||||
return 'imagick';
|
||||
}
|
||||
} catch (Throwable) {
|
||||
// Imagick ohne PDF-Unterstützung
|
||||
}
|
||||
|
||||
return null;
|
||||
}
|
||||
|
||||
/**
|
||||
* @return list<string> Bilddateien der Seiten
|
||||
*/
|
||||
private function rasterize(string $pdf, string $dir): array
|
||||
{
|
||||
if ($this->rasterizer(true) === 'pdftoppm') {
|
||||
$this->run(['pdftoppm', '-r', '300', '-gray', '-png', '-l', (string) self::MAX_PAGES, $pdf, $dir.'/seite']);
|
||||
} else {
|
||||
$imagick = new Imagick;
|
||||
$imagick->setResolution(300, 300);
|
||||
$imagick->readImage($pdf.'[0-'.(self::MAX_PAGES - 1).']');
|
||||
|
||||
foreach ($imagick as $index => $page) {
|
||||
$page->setImageFormat('png');
|
||||
$page->writeImage(sprintf('%s/seite-%03d.png', $dir, $index + 1));
|
||||
}
|
||||
|
||||
$imagick->clear();
|
||||
}
|
||||
|
||||
$images = glob($dir.'/seite*.png') ?: [];
|
||||
natsort($images);
|
||||
|
||||
return array_values($images);
|
||||
}
|
||||
|
||||
/**
|
||||
* Programm ohne Shell aufrufen; null, wenn es nicht läuft.
|
||||
*
|
||||
* @param list<string> $command
|
||||
*/
|
||||
protected function run(array $command): ?string
|
||||
{
|
||||
if (! function_exists('proc_open')) {
|
||||
return null;
|
||||
}
|
||||
|
||||
$process = @proc_open($command, [1 => ['pipe', 'w'], 2 => ['pipe', 'w']], $pipes);
|
||||
|
||||
if (! is_resource($process)) {
|
||||
return null;
|
||||
}
|
||||
|
||||
$output = stream_get_contents($pipes[1]).stream_get_contents($pipes[2]);
|
||||
fclose($pipes[1]);
|
||||
fclose($pipes[2]);
|
||||
|
||||
return proc_close($process) === 0 ? (string) $output : null;
|
||||
}
|
||||
}
|
||||
120
plugins-core/qm/src/Text/TextExtractor.php
Normal file
120
plugins-core/qm/src/Text/TextExtractor.php
Normal file
|
|
@ -0,0 +1,120 @@
|
|||
<?php
|
||||
|
||||
namespace Inhaus\Plugins\Qm\Text;
|
||||
|
||||
use Smalot\PdfParser\Parser;
|
||||
use Throwable;
|
||||
use ZipArchive;
|
||||
|
||||
/**
|
||||
* Text aus Dokumenten für die Volltextsuche – in reinem PHP, läuft also auch auf
|
||||
* Shared Hosting. Eingescannte PDFs ohne Textebene liefern nichts; die übernimmt
|
||||
* (falls eingerichtet) die Texterkennung (Ocr).
|
||||
*/
|
||||
class TextExtractor
|
||||
{
|
||||
/** Mehr Text braucht die Suche nicht. */
|
||||
public const MAX_CHARS = 2_000_000;
|
||||
|
||||
/**
|
||||
* @return array{text: ?string, status: string} status: ok | empty | unsupported | failed
|
||||
*/
|
||||
public function extract(string $path, string $extension): array
|
||||
{
|
||||
try {
|
||||
$text = match (strtolower($extension)) {
|
||||
'pdf' => $this->pdf($path),
|
||||
'docx' => $this->xmlInZip($path, ['word/document.xml', 'word/header1.xml', 'word/footer1.xml']),
|
||||
'xlsx' => $this->xmlInZip($path, ['xl/sharedStrings.xml']),
|
||||
'pptx' => $this->xmlInZip($path, $this->zipEntries($path, '#^ppt/slides/slide\d+\.xml$#')),
|
||||
'odt', 'ods', 'odp' => $this->xmlInZip($path, ['content.xml']),
|
||||
'txt', 'csv' => (string) file_get_contents($path, length: self::MAX_CHARS),
|
||||
default => null,
|
||||
};
|
||||
} catch (Throwable $e) {
|
||||
logger()->warning("QM: Text aus {$path} nicht lesbar: ".$e->getMessage());
|
||||
|
||||
return ['text' => null, 'status' => 'failed'];
|
||||
}
|
||||
|
||||
if ($text === null) {
|
||||
return ['text' => null, 'status' => 'unsupported'];
|
||||
}
|
||||
|
||||
$text = self::normalize($text);
|
||||
|
||||
return ['text' => $text === '' ? null : $text, 'status' => $text === '' ? 'empty' : 'ok'];
|
||||
}
|
||||
|
||||
public static function normalize(string $text): string
|
||||
{
|
||||
$text = mb_convert_encoding($text, 'UTF-8', 'UTF-8');
|
||||
$text = preg_replace('/[^\P{C}\n\t]+/u', ' ', $text) ?? '';
|
||||
$text = preg_replace('/[ \t\x{00A0}]+/u', ' ', $text) ?? '';
|
||||
$text = preg_replace('/\s*\n\s*/', "\n", $text) ?? '';
|
||||
|
||||
return mb_substr(trim($text), 0, self::MAX_CHARS);
|
||||
}
|
||||
|
||||
private function pdf(string $path): string
|
||||
{
|
||||
$parser = new Parser;
|
||||
|
||||
return $parser->parseFile($path)->getText();
|
||||
}
|
||||
|
||||
/**
|
||||
* @param list<string> $entries
|
||||
*/
|
||||
private function xmlInZip(string $path, array $entries): string
|
||||
{
|
||||
$zip = new ZipArchive;
|
||||
|
||||
if ($zip->open($path) !== true) {
|
||||
throw new \RuntimeException('Keine gültige Office-Datei.');
|
||||
}
|
||||
|
||||
$text = '';
|
||||
|
||||
foreach ($entries as $entry) {
|
||||
$xml = $zip->getFromName($entry);
|
||||
|
||||
if ($xml !== false) {
|
||||
// Absätze und Zellen als Zeilen, dann Tags entfernen.
|
||||
$xml = preg_replace('#</(w:p|a:p|text:p|text:h|si|table:table-cell)>#', "\n", $xml);
|
||||
$text .= html_entity_decode(strip_tags((string) $xml), ENT_QUOTES | ENT_XML1, 'UTF-8')."\n";
|
||||
}
|
||||
}
|
||||
|
||||
$zip->close();
|
||||
|
||||
return $text;
|
||||
}
|
||||
|
||||
/**
|
||||
* @return list<string>
|
||||
*/
|
||||
private function zipEntries(string $path, string $pattern): array
|
||||
{
|
||||
$zip = new ZipArchive;
|
||||
|
||||
if ($zip->open($path) !== true) {
|
||||
return [];
|
||||
}
|
||||
|
||||
$names = [];
|
||||
|
||||
for ($i = 0; $i < $zip->numFiles; $i++) {
|
||||
$name = (string) $zip->getNameIndex($i);
|
||||
|
||||
if (preg_match($pattern, $name)) {
|
||||
$names[] = $name;
|
||||
}
|
||||
}
|
||||
|
||||
$zip->close();
|
||||
natsort($names);
|
||||
|
||||
return array_values($names);
|
||||
}
|
||||
}
|
||||
Loading…
Add table
Add a link
Reference in a new issue