Guard against figs wrapped in paragraphs

This is an edge case (in fact, I'm not entirely sure how to reproduce
this from Jupyter Book, but an author's managed it, so), but figures wrapped in <p> tags does cause issues in Atlas. This adds a simple guard against that along with an associated test for coverage.

Author: delfanbaum

jupyter_book_to_htmlbook/figure_processing.py

@@ -8,6 +8,10 @@ def process_figures(chapter):
     """
     figures = chapter.find_all("figure")
     for figure in figures:
+
+        if figure.parent.name == "p":
+            figure.parent.unwrap()
+
         # clean anything extraneous, if extant
         if figure.find_all('a', class_="headerlink") != []:
             for anchor in figure.find_all('a', class_="headerlink"):

tests/test_figure_processing.py

@@ -202,6 +202,36 @@ def test_no_caption_number(self):
         assert result.find("img").get("style") is None
         assert not result.find("span", class_="caption-number")
 
+    def test_fig_wrapped_in_p_is_unwrapped(self):
+        """
+        Edge case where sometimes figures end up wrapped inside <p> tags,
+        causing issues with rendering in Atlas. We should ensure that all
+        figures -- formal or informal -- do not end up wrapped inside
+        paragraphs
+        """
+        fig = """<p><figure class="align-default" id="example-fig">
+<a class="reference internal image-reference" href="images/flower.png">
+<img alt="images/flower.png" src="../_images/flower.png"
+style="height: 150px;" /></a>
+<figcaption>
+<p><span class="caption-number">Fig. 1 </span><span class="caption-text">
+Here is my figure caption!</span>
+<a class="headerlink" href="#example-fig" title="Permalink to this image">#</a>
+</p></figcaption></p>"""
+        informal_fig = '<p><img alt="Flower" src="flower.png" /></p>'
+        text = f"<section>{fig}{informal_fig}</section>"
+        chapter = Soup(text, "html.parser")
+        process_figures(chapter)
+        process_informal_figs(chapter)
+
+        figs = chapter.find_all("figure")
+        assert len(figs) == 2
+
+        for fig in figs:
+            assert not fig.parent.name == "p"
+
+        assert not chapter.find('p')
+
 
 class TestGeneratedFigureProcessing:
     """