diff --git a/src/ptf/cmds/xml/ckeditor/ckeditor_parser.py b/src/ptf/cmds/xml/ckeditor/ckeditor_parser.py
index 3e783a41ea0e96256e5f0c1472b98aa0dbfc9fe9..5f5d0764b4c6bbcf71cf856f7ea089bc85751ada 100644
--- a/src/ptf/cmds/xml/ckeditor/ckeditor_parser.py
+++ b/src/ptf/cmds/xml/ckeditor/ckeditor_parser.py
@@ -32,6 +32,7 @@ from ptf.cmds.xml.xml_utils import escape, get_xml_from_node, normalize, replace
from ptf.display import resolver
from ptf.utils import convert_index_to_list_type
+
# from ptf.utils import create_interlink_for_citation
@@ -51,7 +52,7 @@ class CkeditorParser:
resolve_entities=True,
)
html_value = kwargs["html_value"].replace("\n\n", "")
- body = f"<body>{replace_html_entities(html_value)}</body>"
+ body = f"<fake_body>{replace_html_entities(html_value)}</fake_body>"
tree = etree.fromstring(body.encode("utf-8"), parser=parser)
else:
tree = kwargs["tree"]
@@ -130,10 +131,18 @@ class CkeditorParser:
xml_text += inner_jats_xml_text
xml_text += "</list>"
- # # JATS requires <list> to be inside <p>
- # parent = node.getparent()
- # if parent is None or parent.tag != "p":
- # xml_text = f"<p>{xml_text}</p>"
+ # JATS requires <list> to be inside some specific elements
+ parent = node.getparent()
+ if parent is not None and parent.tag not in ["answer", "app", "app-group", "bio", "body", "boxed-text",
+ "chem-struct", "disp-quote",
+ "explanation", "fig", "glossary", "legend", "license-p",
+ "list-item", "named-content",
+ "notes", "option", "p", "question", "question-preamble",
+ "ref-list", "sec", "styled-content",
+ "supplementary-material",
+ "table-wrap", "td", "th"]:
+ xml_text = f"<p>{xml_text}</p>"
+
html_text = f"<{node.tag}{ol_attributes}>{inner_html_text}</{node.tag}>"
tex_text = f"<{node.tag}{ol_attributes}>{inner_tex_text}</{node.tag}>"
diff --git a/src/ptf/cmds/xml/jats/jats_parser.py b/src/ptf/cmds/xml/jats/jats_parser.py
index b24b0f7212b0175d8bd45a5bbe172dd06b17bcde..e71a1b54a76db714e9e06a778a4f46f7734a4386 100644
--- a/src/ptf/cmds/xml/jats/jats_parser.py
+++ b/src/ptf/cmds/xml/jats/jats_parser.py
@@ -657,6 +657,19 @@ class JatsBase(XmlParserBase):
def parse_node_with_p(self, node, **kwargs):
tex, html = self.parse_inner_node(node, **kwargs)
+ insert_p = True
+ # JATS forces <list> to be inside <p> (or other elements)
+ # In HTML, you would get an extra unwanted <p> to display a list
+ # Ignore the <p> if it only embeds a <list>
+ if node.text is None:
+ children_count = 0
+ for child in node:
+ if child.tag == "list":
+ insert_p = False
+ children_count += 1
+ if not insert_p and children_count == 1:
+ return tex, html
+
if not self.for_tex_file:
tex = f"<p>{tex}</p>"
diff --git a/src/ptf/tests/test_ckeditor.py b/src/ptf/tests/test_ckeditor.py
index 0a1648ea3e40ab97247515362fd5d20fd6978b0e..5484b1cf458c1ded5f120e7194f230d0f62d97ab 100644
--- a/src/ptf/tests/test_ckeditor.py
+++ b/src/ptf/tests/test_ckeditor.py
@@ -7,7 +7,7 @@ def test_ckeditor_parser():
result = parser.value_xml
assert (
result
- == 'Te<st\xa0<inline-formula><alternatives><tex-math>$x = {-b \\pm \\sqrt{b^2-4ac} \\over 2a}$</tex-math></alternatives></inline-formula> done<list list-type="simple"><list-item><p>Item</p></list-item></list><list list-type="number"><list-item><p>Item 1<break/>New line</p></list-item><list-item><p>\xa0</p></list-item></list>'
+ == 'Te<st\xa0<inline-formula><alternatives><tex-math>$x = {-b \\pm \\sqrt{b^2-4ac} \\over 2a}$</tex-math></alternatives></inline-formula> done<p><list list-type="simple"><list-item><p>Item</p></list-item></list></p><p><list list-type="number"><list-item><p>Item 1<break/>New line</p></list-item><list-item><p>\xa0</p></list-item></list></p>'
)
html_value = r"<strong>Item</strong><em>text</em><sub>i</sub><sup>2</sup><a>title</a>"