Preserve void elements before the truncation point in truncatehtml

src/main/java/com/hubspot/jinjava/lib/filter/TruncateHtmlFilter.java

@@ -8,6 +8,8 @@
 import com.hubspot.jinjava.interpret.TemplateError;
 import com.hubspot.jinjava.lib.fn.Functions;
 import com.hubspot.jinjava.objects.SafeString;
+import java.util.ArrayDeque;
+import java.util.Deque;
 import java.util.Map;
 import java.util.Objects;
 import org.apache.commons.lang3.BooleanUtils;
@@ -52,6 +54,7 @@ public class TruncateHtmlFilter implements AdvancedFilter {
 
   private static final int DEFAULT_TRUNCATE_LENGTH = 255;
   private static final String DEFAULT_END = "...";
+  private static final String DELETE_MARKER_CLASS = "__deleteme";
   private static final String LENGTH_KEY = "length";
   private static final String END_KEY = "end";
   private static final String BREAKWORD_KEY = "breakword";
@@ -148,7 +151,7 @@ public Object filter(Object var, JinjavaInterpreter interpreter, String... args)
         killwords
       );
       dom.select("body").traverse(visitor);
-      dom.select(".__deleteme").remove();
+      dom.select("." + DELETE_MARKER_CLASS).remove();
 
       return dom.select("body").html();
     }
@@ -162,6 +165,8 @@ private static class ContentTruncatingNodeVisitor implements NodeVisitor {
     private int textLen;
     private final String ending;
     private final boolean killwords;
+    private final Deque<Integer> keptVoidElementsAtEntry = new ArrayDeque<>();
+    private int keptVoidElements;
 
     ContentTruncatingNodeVisitor(int maxTextLen, String ending, boolean killwords) {
       this.maxTextLen = maxTextLen;
@@ -188,17 +193,33 @@ public void head(Node node, int depth) {
         } else {
           textLen += textContent.length();
         }
+      } else if (node instanceof Element el) {
+        keptVoidElementsAtEntry.push(keptVoidElements);
+        if (isVoidElement(el)) {
+          if (textLen >= maxTextLen) {
+            el.addClass(DELETE_MARKER_CLASS);
+          } else {
+            keptVoidElements++;
+          }
+        }
       }
     }
 
     @Override
     public void tail(Node node, int depth) {
-      if (node instanceof Element) {
-        Element el = (Element) node;
-        if (StringUtils.isBlank(el.text())) {
-          el.addClass("__deleteme");
+      if (node instanceof Element el) {
+        boolean containsKeptVoidElement =
+          keptVoidElements > keptVoidElementsAtEntry.pop();
+        if (
+          !isVoidElement(el) && !containsKeptVoidElement && StringUtils.isBlank(el.text())
+        ) {
+          el.addClass(DELETE_MARKER_CLASS);
         }
       }
     }
+
+    private static boolean isVoidElement(Element el) {
+      return el.tag().isEmpty();
+    }
   }
 }

src/test/java/com/hubspot/jinjava/lib/filter/TruncateHtmlFilterTest.java

@@ -78,6 +78,85 @@ public void itTakesKwargs() {
       );
   }
 
+  @Test
+  public void itPreservesVoidElementsWhenContentIsUnderTheLimit() {
+    String result = (String) filter.filter(
+      "<p>Hey,<br>Great to hear!</p><p><img src=\"a.png\"></p>",
+      interpreter,
+      "255"
+    );
+    assertThat(result)
+      .isEqualTo("<p>Hey,\n <br>\n Great to hear!</p>\n<p><img src=\"a.png\"></p>");
+  }
+
+  @Test
+  public void itPreservesVoidElementsBeforeTheTruncationPoint() {
+    String result = (String) filter.filter(
+      "<p>Hello<br>world, this is long text</p>",
+      interpreter,
+      "10",
+      "...",
+      "true"
+    );
+    assertThat(result).isEqualTo("<p>Hello\n <br>\n world...</p>");
+  }
+
+  @Test
+  public void itRemovesVoidElementsAfterTheTruncationPoint() {
+    String result = (String) filter.filter(
+      "<p>Hello world</p><p>more<br>text</p><img src=\"a.png\">",
+      interpreter,
+      "5",
+      "...",
+      "true"
+    );
+    assertThat(result).isEqualTo("<p>Hello...</p>");
+  }
+
+  @Test
+  public void itPreservesSelfClosingBrTagsBeforeTheTruncationPoint() {
+    String result = (String) filter.filter(
+      "<p>Hey,<br/>Great to hear!</p>",
+      interpreter,
+      "255"
+    );
+    assertThat(result).isEqualTo("<p>Hey,\n <br>\n Great to hear!</p>");
+  }
+
+  @Test
+  public void itRemovesSelfClosingBrTagsAfterTheTruncationPoint() {
+    String result = (String) filter.filter(
+      "<p>Hello world<br/>more text</p><br/>",
+      interpreter,
+      "5",
+      "...",
+      "true"
+    );
+    assertThat(result).isEqualTo("<p>Hello...</p>");
+  }
+
+  @Test
+  public void itPreservesVoidElementsWithExplicitClosingTagsBeforeTheTruncationPoint() {
+    String result = (String) filter.filter(
+      "<p><img src=\"a.png\" width=\"10\"></img>Great to hear!</p>",
+      interpreter,
+      "255"
+    );
+    assertThat(result).isEqualTo("<p><img src=\"a.png\" width=\"10\">Great to hear!</p>");
+  }
+
+  @Test
+  public void itRemovesVoidElementsWithExplicitClosingTagsAfterTheTruncationPoint() {
+    String result = (String) filter.filter(
+      "<p>Hello world</p><p><img src=\"a.png\" width=\"10\"></img></p>",
+      interpreter,
+      "5",
+      "...",
+      "true"
+    );
+    assertThat(result).isEqualTo("<p>Hello...</p>");
+  }
+
   private static String fixture(String name) {
     try {
       return Resources.toString(Resources.getResource(name), StandardCharsets.UTF_8);