fix: strip images from toc labels (#749)

Signed-off-by: Timothée Mazzucotelli <dev@pawamoy.fr>

python/tests/integration/markdown/test_render.py

@@ -104,3 +104,41 @@ class TestTocCleanup:
         # The child heading must preserve the abbreviation text.
         child = result["toc"][0]["children"][0]
         assert child["content"] == "Using CSS"
+
+    def test_images_stripped_from_toc_content(self) -> None:
+        """Images defined in the page must not appear as <img> in TOC."""
+        result = render(
+            content="# ![icon](../../images/system-32.png) System\n",
+            path="index.md",
+            url="/",
+        )
+
+        # Sanity-check: the rendered page body must contain <img>.
+        assert "<img" in result["content"]
+
+        # The TOC must contain exactly one top-level entry.
+        assert len(result["toc"]) == 1
+        heading = result["toc"][0]
+
+        # The TOC content must be plain text – no <img> tags.
+        assert "<img" not in heading["content"]
+        assert heading["content"].strip() == "System"
+
+    def test_images_stripped_from_nested_toc(self) -> None:
+        """Image stripping must apply at every nesting level."""
+        result = render(
+            content=(
+                "# Top level\n\n## ![icon](../../images/system-32.png) System\n"
+            ),
+            path="index.md",
+            url="/",
+        )
+
+        # The rendered page body must contain <img>.
+        assert "<img" in result["content"]
+
+        # Collect content from all TOC levels.
+        all_contents = _toc_contents(result["toc"])
+
+        # No level should contain <img>.
+        assert all("<img" not in c for c in all_contents)

python/tests/unit/markdown/test_render.py

@@ -80,13 +80,30 @@ class TestCleanupTocLabel:
                 "\nAbbr\n",
                 id="multiline_abbr",
             ),
+            # Images ------------------------------------------------------
+            pytest.param(
+                '<img src="image.png" alt="Image">',
+                "",
+                id="img_tag",
+            ),
+            pytest.param(
+                '<img src="image.png" alt="Image">Text',
+                "Text",
+                id="img_tag_with_text_after",
+            ),
+            pytest.param(
+                'Text<img src="image.png" alt="Image">',
+                "Text",
+                id="img_tag_with_text_before",
+            ),
             # Combined and passthrough ------------------------------------
             pytest.param(
                 '<a href="#x">Intro to '
                 '<abbr title="HyperText Markup Language">HTML</abbr>'
+                '<img src="image.png" alt="Image">'
                 "</a>",
                 "Intro to HTML",
-                id="links_and_abbreviations",
+                id="combined_and_passthrough",
             ),
             pytest.param(
                 "Just plain text",

python/zensical/markdown/render.py

@@ -180,4 +180,6 @@ def _cleanup_toc_label(html: str) -> str:
     html = re.sub(r"<a\s+[^>]+>(.*?)</a>", r"\1", html, flags=re.DOTALL)
     # Remove abbreviations
     html = re.sub(r"<abbr\s+[^>]+>(.*?)</abbr>", r"\1", html, flags=re.DOTALL)
+    # Remove images
+    html = re.sub(r"<img\s+[^>]+>", "", html, flags=re.DOTALL)
     return html  # noqa: RET504