Fix bugs, convert thumbnail to png, add context menu

backend/app/services/storage.py

@@ -0,0 +1,53 @@
+"""Document storage service."""
+
+import shutil
+import uuid
+import pymupdf
+from pathlib import Path
+from fastapi import UploadFile, HTTPException
+
+from app.config import settings
+
+def validate_pdf(filepath: Path) -> None:
+    """Validate PDF magic bytes and PyMuPDF openability."""
+    with open(filepath, "rb") as f:
+        header = f.read(5)
+        if header != b"%PDF-":
+            raise ValueError("Not a valid PDF file (missing magic bytes)")
+
+    try:
+        doc = pymupdf.open(filepath)
+        doc.close()
+    except Exception as e:
+        raise ValueError(f"Failed to open PDF with PyMuPDF: {e}")
+
+async def save_upload_file(upload_file: UploadFile) -> tuple[str, int, int]:
+    """Save an uploaded file to disk and validate it."""
+    file_id = str(uuid.uuid4())
+    filepath = settings.PDF_STORAGE_PATH / f"{file_id}.pdf"
+
+    try:
+        with open(filepath, "wb") as f:
+            shutil.copyfileobj(upload_file.file, f)
+    except Exception as e:
+        if filepath.exists():
+            filepath.unlink()
+        raise HTTPException(status_code=500, detail="Failed to save file")
+
+    page_count = 0
+    try:
+        doc = pymupdf.open(filepath)
+        page_count = len(doc)
+        doc.close()
+    except Exception as e:
+        filepath.unlink()
+        raise HTTPException(status_code=400, detail=f"Failed to open PDF with PyMuPDF: {e}")
+    
+    size = filepath.stat().st_size
+    return file_id, size, page_count
+
+def delete_pdf_file(file_id: str) -> None:
+    """Delete a PDF file from storage."""
+    filepath = settings.PDF_STORAGE_PATH / f"{file_id}.pdf"
+    if filepath.exists():
+        filepath.unlink()

backend/app/services/thumbnails.py

@@ -0,0 +1,38 @@
+"""Thumbnail generation service."""
+
+import pymupdf
+from pathlib import Path
+
+from app.config import settings
+
+def generate_thumbnail(file_id: str) -> None:
+    """Generate a PNG thumbnail for the first page of a PDF."""
+    pdf_path = settings.PDF_STORAGE_PATH / f"{file_id}.pdf"
+    thumb_path = settings.THUMBNAILS_PATH / f"{file_id}.png"
+    
+    if not pdf_path.exists():
+        return
+        
+    try:
+        doc = pymupdf.open(pdf_path)
+        if len(doc) > 0:
+            page = doc[0]
+            # Zoom to approximately 600px width
+            rect = page.rect
+            zoom = 600.0 / rect.width if rect.width > 0 else 1.0
+            mat = pymupdf.Matrix(zoom, zoom)
+            
+            # Render pixmap
+            pix = page.get_pixmap(matrix=mat, alpha=False)
+            
+            # Save as PNG
+            pix.save(thumb_path, output="png")
+        doc.close()
+    except Exception as e:
+        print(f"Failed to generate thumbnail for {file_id}: {e}")
+
+def delete_thumbnail(file_id: str) -> None:
+    """Delete a thumbnail file."""
+    thumb_path = settings.THUMBNAILS_PATH / f"{file_id}.png"
+    if thumb_path.exists():
+        thumb_path.unlink()

backend/app/services/trash_sweep.py

@@ -0,0 +1,48 @@
+"""Background tasks for sweeping trash and old versions."""
+
+from datetime import datetime, timedelta, timezone
+from sqlalchemy.orm import Session
+
+from app.db import get_db
+from app.models.document import Document
+from app.models.annotation_state import AnnotationState
+from app.models.version import Version
+from app.services.storage import delete_pdf_file
+from app.services.thumbnails import delete_thumbnail
+from app.config import settings
+
+def sweep_trash(db: Session) -> None:
+    """Permanently delete documents that have been in the trash past the retention period."""
+    cutoff = datetime.now(timezone.utc) - timedelta(days=settings.TRASH_RETENTION_DAYS)
+    cutoff_iso = cutoff.isoformat()
+    
+    docs_to_delete = db.query(Document).filter(
+        Document.in_trash == True,
+        Document.deleted_at <= cutoff_iso
+    ).all()
+    
+    for doc in docs_to_delete:
+        # Delete files from disk
+        delete_pdf_file(doc.id)
+        delete_thumbnail(doc.id)
+        
+        # Delete DB associations
+        db.query(AnnotationState).filter(AnnotationState.document_id == doc.id).delete()
+        db.query(Version).filter(Version.document_id == doc.id).delete()
+        
+        # Delete document record
+        db.delete(doc)
+    
+    db.commit()
+
+def prune_auto_versions(db: Session) -> None:
+    """Delete automatic versions older than the retention period."""
+    cutoff = datetime.now(timezone.utc) - timedelta(days=settings.AUTO_VERSION_RETENTION_DAYS)
+    cutoff_iso = cutoff.isoformat()
+    
+    db.query(Version).filter(
+        Version.kind == "auto",
+        Version.created_at <= cutoff_iso
+    ).delete()
+    
+    db.commit()