Fix #3: Old cover shown after updating book cover

Caddy is set up to serve static files with aggressive caching, which is
great for performance but can cause issues when updating book covers. To
ensure users see the updated cover immediately, we need implement a
cache-busting strategy. This involves generating a unique filename for
each cover image based on its content, using a hash of the image data.
By doing this, when a cover is updated, it will have a new filename,
prompting browsers to fetch the new image instead of using the cached
version.

src/hxbooks/library.py

@@ -5,6 +5,7 @@ Clean service layer for book management, reading tracking, and wishlist operatio
 Separated from web interface concerns to enable both CLI and web access.
 """
 
+import hashlib
 import logging
 from collections import defaultdict
 from collections.abc import Sequence
@@ -1013,28 +1014,23 @@ def download_book_cover(book: Book, image_url: str) -> bool:
                 return False
 
             # Load image directly from file
-            with Image.open(source_path) as image:
-                processed_image = _process_cover_image(image)
-
-                # Generate filename
-                extension = ".jpg"  # Always save as JPEG
-                filename = f"book_{book.id}{extension}"
-                cover_path = covers_dir / filename
-
-                # Save processed image
-                processed_image.save(cover_path, "JPEG", quality=85)
+            source = source_path
         else:
             # Handle HTTP(S) URLs
             response = requests.get(image_url, timeout=10, stream=True)
             response.raise_for_status()
 
             # Load image from response content
-            with Image.open(response.raw) as image:
+            source = response.raw
+
+        with Image.open(source) as image:
             processed_image = _process_cover_image(image)
 
             # Generate filename
             extension = ".jpg"  # Always save as JPEG
-                filename = f"book_{book.id}{extension}"
+            # Hash the image to create a unique filename based on content
+            image_hash = hashlib.md5(processed_image.tobytes()).hexdigest()
+            filename = f"book_{book.id}_{image_hash}{extension}"
             cover_path = covers_dir / filename
 
             # Save processed image