backend/src/main/java/org/booklore/service/metadata/parser/AmazonBookParser.java (1)

886-902: Optional: simplify with pattern matching.

Given Java 25 with --enable-preview is the build target for this codebase, this helper could collapse to a single boolean expression using pattern matching for instanceof, eliminating the explicit casts.

♻️ Proposed simplification

-    private boolean isWhitespaceNode(Node node) {
-        if (node instanceof TextNode) {
-            if (((TextNode) node).isBlank()) {
-                return true;
-            }
-        }
-
-        if (node instanceof Element) {
-            Element element = (Element) node;
-
-            if ("br".equals(element.tagName())) {
-                return true;
-            }
-        }
-
-        return false;
-    }
+    private boolean isWhitespaceNode(Node node) {
+        if (node instanceof TextNode textNode) {
+            return textNode.isBlank();
+        }
+        if (node instanceof Element element) {
+            return "br".equals(element.tagName());
+        }
+        return false;
+    }

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In
`@backend/src/main/java/org/booklore/service/metadata/parser/AmazonBookParser.java`
around lines 886 - 902, Replace the current body of isWhitespaceNode with a
single boolean expression using Java 25 pattern-matching instanceof: return
(node instanceof TextNode t && t.isBlank()) || (node instanceof Element e &&
"br".equals(e.tagName())); this preserves the existing checks (TextNode.isBlank
and Element.tagName() == "br") while removing explicit casts and nested ifs.

backend/src/test/java/org/booklore/service/metadata/parser/AmazonBookParserTest.java (1)

197-220: LGTM — test exercises the new DOM-based cleanup end-to-end.

The HTML input intentionally mixes <br />, <br >, and <br> with surrounding whitespace and embedded <div> blocks, so this verifies both the consecutive-<br> collapsing (more than 2 → 2) and the new leading/trailing whitespace+<br> trim via isWhitespaceNode. Routing Jsoup.parse(anyString()) to the real method (Line 214) is necessary because Jsoup is statically mocked in setup() and cleanDescriptionHtml invokes Jsoup.parse internally.

One small nit: consider asserting on metadata.getDescription() directly (without the \n strip) so any unintended newline reintroduction is caught — body().html().trim() already produces a deterministic, newline-free output for this fixture.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In
`@backend/src/test/java/org/booklore/service/metadata/parser/AmazonBookParserTest.java`
around lines 197 - 220, In AmazonBookParserTest.test
fetchTopMetadata_cleansLineBreaks (method fetchTopMetadata_cleansLineBreaks),
remove the manual newline strip on metadata.getDescription() and assert the
description value directly; specifically replace the current assertEquals that
compares actual (built by metadata.getDescription().replace("\n", "")) with a
direct assertEquals(expected, metadata.getDescription()) so the test will catch
any unintended newline reintroduction by cleanDescriptionHtml/Jsoup.parse.