Show a list item once when its text is wrapped in a paragraph

zeeguu/core/model/article.py

@@ -427,7 +427,7 @@ def create_article_fragments(self, session):
         This preserves the structure and formatting from the original article.
         """
         from zeeguu.core.model.article_fragment import ArticleFragment
-        from bs4 import BeautifulSoup
+        from bs4 import BeautifulSoup, Tag
 
         # Get HTML content - use htmlContent if available, otherwise fall back to plain text
         html_content = getattr(self, "htmlContent", None) or self.source.get_content()
@@ -444,30 +444,50 @@ def create_article_fragments(self, session):
         # Note: We skip ul/ol containers to avoid duplication, only process individual li items
         # Note: We skip inline elements like strong, em here as they should be preserved within their parent blocks
         block_elements = ["p", "h1", "h2", "h3", "h4", "h5", "h6", "li", "blockquote"]
+        text_blocks = [b for b in block_elements if b not in ("li", "blockquote")]
+
+        # A list item may hold its text directly (<li>X</li>) or in blocks
+        # (<li><p>X</p></li>, the shape the teacher editor saves; or a heading
+        # followed by paragraphs). In the latter case the blocks are emitted one
+        # by one, and the first of them carries the bullet.
+        bullet_carriers = set()
 
         for element in soup.find_all(block_elements):
             # Skip blockquote containers - we'll process their paragraph children instead
             if element.name == "blockquote":
                 continue
 
-            # For paragraphs inside blockquotes, use special formatting to indicate they're part of a quote
-            if element.name == "p" and element.find_parent("blockquote"):
-                tag_name = "blockquote"
-                text_content = element.get_text().strip()
-            # For list items, get direct text content (not nested lists)
-            elif element.name == "li":
-                # Get only direct text content, excluding nested ul/ol
+            if element.name == "li":
+                own_blocks = [
+                    b
+                    for b in element.find_all(text_blocks)
+                    if b.find_parent("li") is element
+                ]
+                own_block_ids = {id(b) for b in own_blocks}
+                # Direct text content, excluding nested ul/ol and the blocks
                 text_parts = []
                 for content in element.contents:
-                    if hasattr(content, "get_text"):
-                        # Skip nested lists
-                        if content.name not in ["ul", "ol"]:
-                            text_parts.append(content.get_text().strip())
-                    else:
-                        # Direct text node
+                    if not isinstance(content, Tag):
                         text_parts.append(str(content).strip())
+                    elif content.name in ["ul", "ol"]:
+                        continue
+                    elif id(content) in own_block_ids or any(
+                        id(b) in own_block_ids for b in content.find_all(text_blocks)
+                    ):
+                        continue
+                    else:
+                        text_parts.append(content.get_text().strip())
                 text_content = " ".join(text_parts).strip()
                 tag_name = element.name
+                if not text_content and own_blocks:
+                    bullet_carriers.add(id(own_blocks[0]))
+            elif id(element) in bullet_carriers:
+                text_content = element.get_text().strip()
+                tag_name = "li"
+            # For paragraphs inside blockquotes, use special formatting to indicate they're part of a quote
+            elif element.name == "p" and element.find_parent("blockquote"):
+                tag_name = "blockquote"
+                text_content = element.get_text().strip()
             else:
                 # For other elements, preserve some inline formatting by getting inner HTML
                 # and then extracting text, but keep track of formatting

zeeguu/core/test/test_article_fragments.py

@@ -0,0 +1,59 @@
+from unittest import TestCase
+
+import zeeguu.core
+from zeeguu.core.model.article_fragment import ArticleFragment
+from zeeguu.core.test.model_test_mixin import ModelTestMixIn
+from zeeguu.core.test.rules.article_rule import ArticleRule
+
+db_session = zeeguu.core.model.db.session
+
+
+class ArticleFragmentsTest(ModelTestMixIn, TestCase):
+    """Each block of the article's HTML becomes exactly one reader fragment."""
+
+    def _fragments(self, html):
+        article = ArticleRule().article
+        ArticleFragment.query.filter_by(article_id=article.id).delete()
+        article.htmlContent = html
+        article.create_article_fragments(db_session)
+        db_session.commit()
+        return [
+            (f.formatting, f.text.content)
+            for f in ArticleFragment.get_all_article_fragments_in_order(article.id)
+        ]
+
+    def test_paragraph_inside_list_item_is_not_repeated(self):
+        # The shape the teacher editor (TipTap) saves every list item in
+        html = "<p>Vessels:</p><ul><li><p>Container ships</p></li><li><p>Tankers</p></li></ul>"
+        self.assertEqual(
+            [("p", "Vessels:"), ("li", "Container ships"), ("li", "Tankers")],
+            self._fragments(html),
+        )
+
+    def test_nested_list_items_each_appear_once(self):
+        html = "<ul><li><p>Ships</p><ul><li><p>Tankers</p></li></ul></li></ul>"
+        self.assertEqual(
+            [("li", "Ships"), ("li", "Tankers")],
+            self._fragments(html),
+        )
+
+    def test_list_item_with_several_blocks_keeps_them_apart(self):
+        html = "<ul><li><h3>Container ships</h3><p>Carry boxes.</p><p>Largest class.</p></li></ul>"
+        self.assertEqual(
+            [("li", "Container ships"), ("p", "Carry boxes."), ("p", "Largest class.")],
+            self._fragments(html),
+        )
+
+    def test_loose_text_before_blocks_carries_the_bullet(self):
+        html = "<ul><li><strong>Tankers</strong><p>Carry oil.</p></li></ul>"
+        self.assertEqual(
+            [("li", "Tankers"), ("p", "Carry oil.")],
+            self._fragments(html),
+        )
+
+    def test_plain_list_items_and_quotes_unchanged(self):
+        html = "<ol><li>One</li></ol><blockquote><p>Quoted</p></blockquote>"
+        self.assertEqual(
+            [("li", "One"), ("blockquote", "Quoted")],
+            self._fragments(html),
+        )