Fix run break order and manual page breaks

docx2everything/converters/markdown_converter.py

@@ -83,32 +83,30 @@ def parse_run_to_markdown(r_elem, hyperlinks=None, images=None, img_dir=None, zi
     text = ''
     rPr = r_elem.find(qn('w:rPr'))
 
-    # Extract text from runs
-    for t_elem in r_elem.findall(qn('w:t')):
-        if t_elem.text:
-            text += t_elem.text
-
-    # Handle tabs and breaks
-    for tab in r_elem.findall(qn('w:tab')):
-        text += '    '  # Convert tab to 4 spaces
-    for br in r_elem.findall(qn('w:br')):
-        text += '\n'
-
-    # Handle footnote references
-    for footnote_ref in r_elem.findall(qn('w:footnoteReference')):
-        footnote_id = footnote_ref.get('{http://schemas.openxmlformats.org/wordprocessingml/2006/main}id')
-        if footnotes and footnote_id in footnotes:
-            text += f'[^{footnote_id}]'
-        else:
-            text += f'[^{footnote_id}]'
-
-    # Handle endnote references
-    for endnote_ref in r_elem.findall(qn('w:endnoteReference')):
-        endnote_id = endnote_ref.get('{http://schemas.openxmlformats.org/wordprocessingml/2006/main}id')
-        if endnotes and endnote_id in endnotes:
-            text += f'[^{endnote_id}]'
-        else:
-            text += f'[^{endnote_id}]'
+    # Extract run content in document order.
+    for child in r_elem:
+        if child.tag == qn('w:t') and child.text:
+            text += child.text
+        elif child.tag == qn('w:tab'):
+            text += '    '  # Convert tab to 4 spaces
+        elif child.tag == qn('w:br'):
+            break_type = child.get('{http://schemas.openxmlformats.org/wordprocessingml/2006/main}type')
+            if break_type == 'page':
+                text += '\n\n<!-- Page Break -->\n\n'
+            else:
+                text += '\n'
+        elif child.tag == qn('w:footnoteReference'):
+            footnote_id = child.get('{http://schemas.openxmlformats.org/wordprocessingml/2006/main}id')
+            if footnotes and footnote_id in footnotes:
+                text += f'[^{footnote_id}]'
+            else:
+                text += f'[^{footnote_id}]'
+        elif child.tag == qn('w:endnoteReference'):
+            endnote_id = child.get('{http://schemas.openxmlformats.org/wordprocessingml/2006/main}id')
+            if endnotes and endnote_id in endnotes:
+                text += f'[^{endnote_id}]'
+            else:
+                text += f'[^{endnote_id}]'
 
     if not text:
         return ''

tests/test_markdown_manual_page_break.py

@@ -0,0 +1,42 @@
+import xml.etree.ElementTree as ET
+
+from docx2everything.converters.markdown_converter import parse_run_to_markdown
+
+
+W_NS = "http://schemas.openxmlformats.org/wordprocessingml/2006/main"
+
+
+def run_with_break(break_type=None):
+    type_attr = "" if break_type is None else f' w:type="{break_type}"'
+    xml = f"""
+    <w:r xmlns:w="{W_NS}">
+      <w:t>Before</w:t>
+      <w:br{type_attr}/>
+      <w:t>After</w:t>
+    </w:r>
+    """
+    return ET.fromstring(xml)
+
+
+def test_page_break_run_outputs_page_break_marker():
+    run = run_with_break("page")
+
+    markdown = parse_run_to_markdown(run)
+
+    assert markdown == "Before\n\n<!-- Page Break -->\n\nAfter"
+
+
+def test_regular_break_run_outputs_newline():
+    run = run_with_break()
+
+    markdown = parse_run_to_markdown(run)
+
+    assert markdown == "Before\nAfter"
+
+
+def test_text_wrapping_break_run_outputs_newline():
+    run = run_with_break("textWrapping")
+
+    markdown = parse_run_to_markdown(run)
+
+    assert markdown == "Before\nAfter"