Fine tune filtering.

Author: dereuromark

src/Filter/ProfileFilter.php

@@ -7,10 +7,16 @@
 use Djot\Exception\ProfileViolationException;
 use Djot\LinkPolicy;
 use Djot\Node\Block\BlockNode;
+use Djot\Node\Block\BlockQuote;
 use Djot\Node\Block\Paragraph;
+use Djot\Node\Block\Table;
+use Djot\Node\Block\TableRow;
 use Djot\Node\Document;
+use Djot\Node\Inline\FootnoteRef;
+use Djot\Node\Inline\HardBreak;
 use Djot\Node\Inline\Image;
 use Djot\Node\Inline\Link;
+use Djot\Node\Inline\Symbol;
 use Djot\Node\Inline\Text;
 use Djot\Node\Node;
 use Djot\Profile;
@@ -185,7 +191,7 @@ protected function convertToText(Node $node, Node $parent): void
         // For inline nodes, just replace with text
         if ($node instanceof BlockNode) {
             $paragraph = new Paragraph();
-            $paragraph->appendChild(new Text($textContent));
+            $this->appendTextWithBreaks($paragraph, $textContent);
             $parent->replaceChildNode($node, $paragraph);
         } else {
             // Inline node - replace with text
@@ -194,6 +200,25 @@ protected function convertToText(Node $node, Node $parent): void
         }
     }
 
+    /**
+     * Append text content to a node, converting newlines to HardBreak nodes
+     */
+    protected function appendTextWithBreaks(Node $parent, string $content): void
+    {
+        $lines = explode("\n", $content);
+        $lastIndex = count($lines) - 1;
+
+        foreach ($lines as $index => $line) {
+            if ($line !== '') {
+                $parent->appendChild(new Text($line));
+            }
+            // Add line break between lines (not after the last line)
+            if ($index < $lastIndex) {
+                $parent->appendChild(new HardBreak());
+            }
+        }
+    }
+
     /**
      * Remove empty container nodes (list items, paragraphs with no content, empty lists)
      */
@@ -264,6 +289,45 @@ protected function extractTextContent(Node $node): string
             return $text;
         }
 
+        // Special handling for tables - preserve row structure
+        if ($node instanceof Table) {
+            $rows = [];
+            foreach ($node->getChildren() as $row) {
+                if ($row instanceof TableRow) {
+                    $cells = [];
+                    foreach ($row->getChildren() as $cell) {
+                        $cells[] = $this->extractTextContent($cell);
+                    }
+                    $rows[] = implode(' ', $cells);
+                }
+            }
+
+            return implode("\n", $rows);
+        }
+
+        // Special handling for blockquotes - preserve paragraph structure
+        if ($node instanceof BlockQuote) {
+            $paragraphs = [];
+            foreach ($node->getChildren() as $child) {
+                $text = $this->extractTextContent($child);
+                if ($text !== '') {
+                    $paragraphs[] = $text;
+                }
+            }
+
+            return implode("\n", $paragraphs);
+        }
+
+        // Special handling for symbols - use the symbol name
+        if ($node instanceof Symbol) {
+            return ':' . $node->getName() . ':';
+        }
+
+        // Special handling for footnote references - use the label
+        if ($node instanceof FootnoteRef) {
+            return '[^' . $node->getLabel() . ']';
+        }
+
         if ($node instanceof Text) {
             return $node->getContent();
         }

tests/TestCase/ProfileTest.php

@@ -802,6 +802,30 @@ public function testCodeBlockContentPreservedAsText(): void
         $this->assertStringContainsString('code block content', $html);
     }
 
+    public function testTableConvertedToStructuredText(): void
+    {
+        $profile = Profile::comment();
+        $converter = new DjotConverter(profile: $profile);
+
+        $djot = <<<'DJOT'
+| Name | Type   |
+|------|--------|
+| Djot | Markup |
+| PHP  | Code   |
+DJOT;
+
+        $html = $converter->convert($djot);
+
+        // Table tag should not appear
+        $this->assertStringNotContainsString('<table>', $html);
+        // Content should be preserved with row structure
+        $this->assertStringContainsString('Name Type', $html);
+        $this->assertStringContainsString('Djot Markup', $html);
+        $this->assertStringContainsString('PHP Code', $html);
+        // Rows should be on separate lines (converted to <br> tags)
+        $this->assertStringContainsString('Name Type<br>', $html);
+    }
+
     // ==================== Whitespace Preservation Tests ====================
 
     public function testFilteredBlocksDoNotRunTogether(): void
@@ -860,4 +884,56 @@ public function testNestedBlockContentPreservesWhitespace(): void
         // Should not run together
         $this->assertStringNotContainsString('First lineSecond line', $html);
     }
+
+    public function testSymbolConvertedToTextRepresentation(): void
+    {
+        $profile = Profile::minimal();
+        $converter = new DjotConverter(profile: $profile);
+
+        $html = $converter->convert('I :heart: this!');
+
+        // Symbol tag should not appear
+        $this->assertStringNotContainsString('<symbol', $html);
+        // Symbol name should be preserved with colons
+        $this->assertStringContainsString(':heart:', $html);
+    }
+
+    public function testFootnoteRefConvertedToTextRepresentation(): void
+    {
+        $profile = Profile::minimal();
+        $converter = new DjotConverter(profile: $profile);
+
+        $djot = <<<'DJOT'
+This has a footnote[^1].
+
+[^1]: The footnote content.
+DJOT;
+
+        $html = $converter->convert($djot);
+
+        // Footnote elements should not appear
+        $this->assertStringNotContainsString('<a href="#fn', $html);
+        // Footnote reference should be converted to text
+        $this->assertStringContainsString('[^1]', $html);
+    }
+
+    public function testBlockquoteWithMultipleParagraphsPreservesLineBreaks(): void
+    {
+        $profile = Profile::minimal();
+        $converter = new DjotConverter(profile: $profile);
+
+        $djot = <<<'DJOT'
+> First paragraph.
+>
+> Second paragraph.
+DJOT;
+
+        $html = $converter->convert($djot);
+
+        $this->assertStringNotContainsString('<blockquote>', $html);
+        $this->assertStringContainsString('First paragraph.', $html);
+        $this->assertStringContainsString('Second paragraph.', $html);
+        // Paragraphs should be separated by line breaks
+        $this->assertStringContainsString('<br>', $html);
+    }
 }