refactor: simplify PDF to image conversion by reading generated files directly from temp directory instead of pre-calculating page count

paperless-backend/src/preprocessing/pdf.service.ts

@@ -40,11 +40,7 @@ export class PdfService {
    * Verwendet einen einzigen Ghostscript-Aufruf mit %d-Platzhalter für alle Seiten.
    */
   async pdfToImages(pdfPath: string, dpi = 200): Promise<string[]> {
-    const pageCount = await this.getPageCount(pdfPath);
-    if (pageCount === 0) return [];
-
     const tmpDir = await fs.mkdtemp(path.join(os.tmpdir(), 'pdf-'));
-    const outputPattern = path.join(tmpDir, 'page-%d.png');
 
     await execFileAsync('gs', [
       '-dNOPAUSE',
@@ -52,22 +48,26 @@ export class PdfService {
       '-dSAFER',
       '-sDEVICE=png16m',
       `-r${dpi}`,
-      `-sOutputFile=${outputPattern}`,
+      `-sOutputFile=${path.join(tmpDir, 'page-%d.png')}`,
       pdfPath,
     ]);
 
-    const images: string[] = [];
+    const entries = await fs.readdir(tmpDir);
-    for (let i = 1; i <= pageCount; i++) {
+    const images = entries
-      const imgPath = path.join(tmpDir, `page-${i}.png`);
+      .filter(f => f.endsWith('.png'))
-      try {
+      .sort((a, b) => {
-        await fs.access(imgPath);
+        const numA = parseInt(a.match(/\d+/)?.[0] ?? '0', 10);
-        images.push(imgPath);
+        const numB = parseInt(b.match(/\d+/)?.[0] ?? '0', 10);
-      } catch {
+        return numA - numB;
-        this.logger.warn(`Ghostscript hat Seite ${i} nicht erstellt: ${imgPath}`);
+      })
-      }
+      .map(f => path.join(tmpDir, f));
+
+    if (images.length === 0) {
+      this.logger.warn(`Ghostscript hat keine Seiten erstellt: ${pdfPath} — Verzeichnisinhalt: [${entries.join(', ')}]`);
+    } else {
+      this.logger.debug(`PDF konvertiert: ${images.length} Seite(n) in ${tmpDir}`);
     }
 
-    this.logger.debug(`PDF konvertiert: ${images.length}/${pageCount} Seite(n) in ${tmpDir}`);
     return images;
   }