fix: early PDF validation, type-aware errors, exact job dedup

- Move PDF page count check before MinIO upload and DB record creation
  to prevent orphaned objects when >500 page PDFs are rejected
- Use job type label (PDF page / Video) in media-worker error messages
  instead of hardcoded "Video processing failed"
- Replace LIKE-based idempotency check with exact JSON match to prevent
  page 1 matching page 11/12/etc substring collisions

backend/routes/pdf.js

@@ -67,8 +67,8 @@ router.post('/:boardId/pdf-pages', async (req, res) => {
       // Check for existing pending/processing hires job for this page
       const { db } = require('../db');
       const existingJob = db.prepare(
-        `SELECT id FROM media_jobs WHERE image_id = ? AND type = 'pdf-hires' AND result_json LIKE ? AND status IN ('queued', 'processing')`
+        `SELECT id FROM media_jobs WHERE image_id = ? AND type = 'pdf-hires' AND result_json = ? AND status IN ('queued', 'processing')`
-      ).get(imageId, `%"pageNumber":${pageNum}%`);
+      ).get(imageId, JSON.stringify({ pageNumber: pageNum }));
 
       if (!existingJob) {
         createMediaJob({
@@ -126,8 +126,8 @@ router.post('/:boardId/pdf-thumbnails', async (req, res) => {
       // Skip if job already pending
       const { db } = require('../db');
       const existingJob = db.prepare(
-        `SELECT id FROM media_jobs WHERE image_id = ? AND type = 'pdf-thumbnail' AND result_json LIKE ? AND status IN ('queued', 'processing')`
+        `SELECT id FROM media_jobs WHERE image_id = ? AND type = 'pdf-thumbnail' AND result_json = ? AND status IN ('queued', 'processing')`
-      ).get(imageId, `%"pageNumber":${pageNum}%`);
+      ).get(imageId, JSON.stringify({ pageNumber: pageNum }));
 
       if (!existingJob) {
         createMediaJob({

backend/routes/upload.js

@@ -131,33 +131,7 @@ router.post('/boards/:boardId/images', upload.single('image'), async (req, res)
     const { buffer, originalname, mimetype, size } = req.file;
     const mediaType = classifyMedia(mimetype);
 
-    const { assetKey, minioPath, width, height, isVideo } = await uploadMedia(board.id, imageId, buffer, mimetype);
+    // PDF: validate page count BEFORE uploading to MinIO or creating DB records
-    const publicUrl = getImageUrl(minioPath);
-
-    // Save image record first (media_jobs FK references images)
-    const image = createImage({
-      id: imageId,
-      boardId: board.id,
-      filename: originalname,
-      mimeType: mimetype,
-      fileSize: size,
-      width,
-      height,
-      minioPath,
-      publicUrl,
-      uploadedBy: req.user.id,
-      assetKey,
-      mediaType,
-    });
-
-    // Enqueue background processing after image record exists
-    let jobId = null;
-    if (isVideo) {
-      jobId = uuidv4();
-      createMediaJob({ id: jobId, imageId, boardId: board.id, type: 'poster' });
-    }
-
-    // PDF: extract page info, create pdf_pages rows, queue thumbnail jobs
     if (mediaType === 'pdf') {
       const { tmpPath, cleanup } = bufferToTempFile(buffer, '.pdf');
       try {
@@ -166,6 +140,25 @@ router.post('/boards/:boardId/images', upload.single('image'), async (req, res)
           return res.status(400).json({ error: 'PDF exceeds 500 page limit' });
         }
 
+        // Validation passed — now upload and create records
+        const { assetKey, minioPath, width, height } = await uploadMedia(board.id, imageId, buffer, mimetype);
+        const publicUrl = getImageUrl(minioPath);
+
+        const image = createImage({
+          id: imageId,
+          boardId: board.id,
+          filename: originalname,
+          mimeType: mimetype,
+          fileSize: size,
+          width,
+          height,
+          minioPath,
+          publicUrl,
+          uploadedBy: req.user.id,
+          assetKey,
+          mediaType,
+        });
+
         updateImagePageCount(imageId, info.pageCount);
 
         // Create pdf_pages rows for all pages
@@ -214,6 +207,32 @@ router.post('/boards/:boardId/images', upload.single('image'), async (req, res)
       }
     }
 
+    const { assetKey, minioPath, width, height, isVideo } = await uploadMedia(board.id, imageId, buffer, mimetype);
+    const publicUrl = getImageUrl(minioPath);
+
+    // Save image record first (media_jobs FK references images)
+    const image = createImage({
+      id: imageId,
+      boardId: board.id,
+      filename: originalname,
+      mimeType: mimetype,
+      fileSize: size,
+      width,
+      height,
+      minioPath,
+      publicUrl,
+      uploadedBy: req.user.id,
+      assetKey,
+      mediaType,
+    });
+
+    // Enqueue background processing after image record exists
+    let jobId = null;
+    if (isVideo) {
+      jobId = uuidv4();
+      createMediaJob({ id: jobId, imageId, boardId: board.id, type: 'poster' });
+    }
+
     return res.status(201).json({
       id: image.id,
       url: publicUrl,

backend/services/media-worker.js

@@ -144,13 +144,14 @@ async function processJob(job) {
     console.error('[media-worker] Job %s failed:', jobId, err.message);
 
     // Classify error for user-facing message
-    let userError = 'Video processing failed';
+    const typeLabel = job.type.startsWith('pdf-') ? 'PDF page' : 'Video';
+    let userError = `${typeLabel} processing failed`;
     if (err.killed || err.signal === 'SIGTERM') {
-      userError = 'Video processing timed out — file may be too large or corrupt';
+      userError = `${typeLabel} processing timed out — file may be too large or corrupt`;
     } else if (err.message?.includes('ENOMEM') || err.message?.includes('Cannot allocate')) {
-      userError = 'Out of memory — video file is too large to process';
+      userError = `Out of memory — file is too large to process`;
     } else if (err.message?.includes('Invalid data')) {
-      userError = 'Invalid or corrupt video file';
+      userError = `Invalid or corrupt ${typeLabel.toLowerCase()} file`;
     }
 
     const attempts = (job.attempts || 0) + 1;