fixes #3477: apoc.load.html does not always report href

neo4j-contrib · Feb 28, 2023 · 4fa02eb · 4fa02eb
1 parent 490c6b5
commit 4fa02eb
Show file tree

Hide file tree

Showing 2 changed files with 30 additions and 2 deletions.
diff --git a/full/src/main/java/apoc/load/LoadHtml.java b/full/src/main/java/apoc/load/LoadHtml.java
@@ -5,6 +5,8 @@
 import apoc.util.MissingDependencyException;
 import apoc.util.FileUtils;
 import java.nio.charset.UnsupportedCharsetException;
+
+import org.apache.commons.lang3.StringUtils;
 import org.jsoup.Jsoup;
 import org.jsoup.nodes.Attribute;
 import org.jsoup.nodes.Document;
@@ -74,7 +76,7 @@ private Stream<MapResult> readHtmlPage(String url, Map<String, String> query, Ma
             }
 
             return Stream.of(new MapResult(output));
-        } catch ( UnsupportedCharsetException e) {
+        } catch (UnsupportedCharsetException e) {
             throw new RuntimeException("Unsupported charset: " + config.getCharset());
         } catch (IllegalArgumentException | ClassCastException e) {
             throw new RuntimeException("Invalid config: " + config);
@@ -137,7 +139,16 @@ private static Map<String, String> getAttributes(Element element) {
                 final String key = attribute.getKey();
                 // with href/src attribute we prepend baseUri path
                 final boolean attributeHasLink = key.equals("href") || key.equals("src");
-                attributes.put(key, attributeHasLink ? element.absUrl(key) : attribute.getValue());
+                String attr = null;
+                if (attributeHasLink) {
+                    attr = element.absUrl(key);
+                    if (StringUtils.isBlank(attr)) {
+                        attr = attribute.getValue();
+                    }
+                } else {
+                    attr = attribute.getValue();
+                }
+                attributes.put(key, attr);
             }
         }
 

diff --git a/full/src/test/java/apoc/load/LoadHtmlTest.java b/full/src/test/java/apoc/load/LoadHtmlTest.java
@@ -3,6 +3,7 @@
 import apoc.ApocSettings;
 import apoc.util.TestUtil;
 import org.apache.commons.lang.exception.ExceptionUtils;
+import org.junit.Assert;
 import org.junit.Before;
 import org.junit.Rule;
 import org.junit.Test;
@@ -353,6 +354,22 @@ public void testQueryWithFailsSilentlyWithLog() {
                 });
     }
 
+    @Test
+    public void testHref() {
+        String query = "CALL apoc.load.html('https://www.amazon.it/', { href: 'div#nav-xshop > a[class=\"nav-a  \"]'}) YIELD value\n" +
+                "UNWIND value.href AS row\n" +
+                "WITH row\n" +
+                "WHERE row.text = 'Bestseller'\n" +
+                "RETURN row ";
+
+        testResult(db, query, Map.of(),
+                result -> {
+                    Map<String, Object> row = (Map<String, Object>) result.next().get("row");
+                    Map<String, Object> attributes = (Map<String, Object>) row.get("attributes");
+                    Assert.assertEquals("/gp/bestsellers/?ref_=nav_cs_bestsellers", attributes.get("href"));
+                });
+    }
+
     @Test
     public void testQueryWithFailsSilentlyWithList() {
         Map<String, Object> query = map("a", "a", "invalid", "invalid", "h6", "h6");