From 3302e783a6c44c2e34ab55d8211210d02e006099 Mon Sep 17 00:00:00 2001
From: Michele Dolfi <dol@zurich.ibm.com>
Date: Fri, 31 Jan 2025 11:22:53 +0100
Subject: [PATCH 01/11] remove un-needed logic

the labels allowlist is checked before

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
---
 docling_core/types/doc/document.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
index 87815319..db5f2ed4 100644
--- a/docling_core/types/doc/document.py
+++ b/docling_core/types/doc/document.py
@@ -2481,7 +2481,7 @@ def _sanitize_text(text: str, do_escape_html=True) -> str:
                 text = f"<li>{_sanitize_text(item.text)}</li>"
                 html_texts.append(text)
 
-            elif isinstance(item, CodeItem) and item.label in labels:
+            elif isinstance(item, CodeItem):
                 text = (
                     "<pre><code>"
                     f"{_sanitize_text(item.text, do_escape_html=False)}"
@@ -2489,7 +2489,7 @@ def _sanitize_text(text: str, do_escape_html=True) -> str:
                 )
                 html_texts.append(text.strip())
 
-            elif isinstance(item, TextItem) and item.label in labels:
+            elif isinstance(item, TextItem):
 
                 text = f"<p>{_sanitize_text(item.text)}</p>"
                 html_texts.append(text.strip())

From de8f4aadd5b4922535b1fae2c0b9520746a406cd Mon Sep 17 00:00:00 2001
From: Michele Dolfi <dol@zurich.ibm.com>
Date: Fri, 31 Jan 2025 11:24:36 +0100
Subject: [PATCH 02/11] textitem cannot have label code

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
---
 docling_core/types/doc/document.py | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
index db5f2ed4..2c7962bd 100644
--- a/docling_core/types/doc/document.py
+++ b/docling_core/types/doc/document.py
@@ -2466,11 +2466,6 @@ def _sanitize_text(text: str, do_escape_html=True) -> str:
                 )
                 html_texts.append(text.strip())
 
-            elif isinstance(item, TextItem) and item.label in [DocItemLabel.CODE]:
-
-                text = f"<pre>{_sanitize_text(item.text, do_escape_html=False)}</pre>"
-                html_texts.append(text)
-
             elif isinstance(item, ListItem):
 
                 text = f"<li>{_sanitize_text(item.text)}</li>"

From d42f37565309fc6f7a56397c00e77ddd89a26ded Mon Sep 17 00:00:00 2001
From: Michele Dolfi <dol@zurich.ibm.com>
Date: Fri, 31 Jan 2025 13:24:13 +0100
Subject: [PATCH 03/11] display formulas with mathml in exported html

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
---
 docling_core/types/doc/document.py            | 25 +++++++++++++++++++
 poetry.lock                                   | 13 +++++++++-
 pyproject.toml                                |  1 +
 .../export/formula_mathml.html                |  5 ++++
 test/test_docling_doc.py                      | 14 +++++++++++
 5 files changed, 57 insertions(+), 1 deletion(-)
 create mode 100644 test/data/docling_document/export/formula_mathml.html

diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
index 2c7962bd..94f37e32 100644
--- a/docling_core/types/doc/document.py
+++ b/docling_core/types/doc/document.py
@@ -16,7 +16,10 @@
 from pathlib import Path
 from typing import Any, Dict, Final, List, Literal, Optional, Tuple, Union
 from urllib.parse import quote, unquote
+from xml.etree.cElementTree import SubElement, tostring
+from xml.sax.saxutils import unescape
 
+import latex2mathml.converter
 import pandas as pd
 import yaml
 from PIL import Image as PILImage
@@ -2347,6 +2350,7 @@ def export_to_html(  # noqa: C901
         to_element: int = sys.maxsize,
         labels: set[DocItemLabel] = DEFAULT_EXPORT_LABELS,
         image_mode: ImageRefMode = ImageRefMode.PLACEHOLDER,
+        formula_to_mathml: bool = False,
         page_no: Optional[int] = None,
         html_lang: str = "en",
         html_head: str = _HTML_DEFAULT_HEAD,
@@ -2466,6 +2470,27 @@ def _sanitize_text(text: str, do_escape_html=True) -> str:
                 )
                 html_texts.append(text.strip())
 
+            elif isinstance(item, TextItem) and item.label in [DocItemLabel.FORMULA]:
+
+                if formula_to_mathml:
+                    # Building a math equation in MathML format
+                    # ref https://www.w3.org/TR/wai-aria-1.1/#math
+                    math_formula = _sanitize_text(item.text, do_escape_html=False)
+                    mathml_element = latex2mathml.converter.convert_to_element(
+                        math_formula, display="block"
+                    )
+                    annotation = SubElement(
+                        mathml_element, "annotation", dict(encoding="TeX")
+                    )
+                    annotation.text = math_formula
+                    mathml = unescape(tostring(mathml_element, encoding="unicode"))
+                    text = f"<div>{mathml}</div>"
+                else:
+                    text = (
+                        f"<pre>{_sanitize_text(item.text, do_escape_html=False)}</pre>"
+                    )
+                html_texts.append(text)
+
             elif isinstance(item, ListItem):
 
                 text = f"<li>{_sanitize_text(item.text)}</li>"
diff --git a/poetry.lock b/poetry.lock
index c039dbe1..79535c36 100644
--- a/poetry.lock
+++ b/poetry.lock
@@ -862,6 +862,17 @@ enabler = ["pytest-enabler (>=2.2)"]
 test = ["pyfakefs", "pytest (>=6,!=8.1.*)"]
 type = ["pygobject-stubs", "pytest-mypy", "shtab", "types-pywin32"]
 
+[[package]]
+name = "latex2mathml"
+version = "3.77.0"
+description = "Pure Python library for LaTeX to MathML conversion"
+optional = false
+python-versions = ">=3.8.1,<4.0.0"
+files = [
+    {file = "latex2mathml-3.77.0-py3-none-any.whl", hash = "sha256:5531e18a2a9eae7c24e257118b6a444cbba253cd27ff3e81f1bd6c41e88e786e"},
+    {file = "latex2mathml-3.77.0.tar.gz", hash = "sha256:e2f501d1878f2e489c3f6f12786bef74c62f712d2770f7f3c837eb20a55d0a1e"},
+]
+
 [[package]]
 name = "markdown-it-py"
 version = "3.0.0"
@@ -2682,4 +2693,4 @@ chunking = ["semchunk", "transformers"]
 [metadata]
 lock-version = "2.0"
 python-versions = "^3.9"
-content-hash = "0475894c9c8ec390cda2f68e75edbf3be6601224090e8a6d826bfc4aa4f36aeb"
+content-hash = "10d96addf225f2632b23451d92cb0641f6724fd9bcd59939b45f0ad86351fb20"
diff --git a/pyproject.toml b/pyproject.toml
index bc11419d..3b2f0aa1 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -59,6 +59,7 @@ typing-extensions = "^4.12.2"
 transformers = { version = "^4.34.0", optional = true }
 semchunk = { version = "^2.2.0", optional = true }
 typer = "^0.12.5"
+latex2mathml = "^3.77.0"
 
 [tool.poetry.extras]
 chunking = ["transformers", "semchunk"]
diff --git a/test/data/docling_document/export/formula_mathml.html b/test/data/docling_document/export/formula_mathml.html
new file mode 100644
index 00000000..58f3435f
--- /dev/null
+++ b/test/data/docling_document/export/formula_mathml.html
@@ -0,0 +1,5 @@
+<!DOCTYPE html>
+<html lang="en">
+
+<div><math xmlns="http://www.w3.org/1998/Math/MathML" display="block"><mrow><mfrac><mrow><mn>1</mn></mrow><mrow><mi>x</mi></mrow></mfrac></mrow><annotation encoding="TeX">\frac{1}{x}</annotation></math></div>
+</html>
\ No newline at end of file
diff --git a/test/test_docling_doc.py b/test/test_docling_doc.py
index c29ae59b..e7bc1412 100644
--- a/test/test_docling_doc.py
+++ b/test/test_docling_doc.py
@@ -661,6 +661,20 @@ def test_version_doc():
     assert doc.version == CURRENT_VERSION
 
 
+def test_formula_mathml():
+    doc = DoclingDocument(name="Dummy")
+    equation = "\\frac{1}{x}"
+    doc.add_text(label=DocItemLabel.FORMULA, text=equation)
+
+    doc_html = doc.export_to_html(formula_to_mathml=True, html_head="")
+
+    gt_html = Path("test/data/docling_document/export/formula_mathml.html").read_text(
+        encoding="utf8"
+    )
+
+    assert doc_html == gt_html
+
+
 def test_docitem_get_image():
     # Prepare the document
     doc = DoclingDocument(name="Dummy")

From a59501e95d4eca82caa2ebabf307c9c82ff238c4 Mon Sep 17 00:00:00 2001
From: Michele Dolfi <dol@zurich.ibm.com>
Date: Fri, 31 Jan 2025 13:25:38 +0100
Subject: [PATCH 04/11] expose argument in save_as_html

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
---
 docling_core/types/doc/document.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
index 94f37e32..83df77bf 100644
--- a/docling_core/types/doc/document.py
+++ b/docling_core/types/doc/document.py
@@ -2285,6 +2285,7 @@ def save_as_html(
         to_element: int = sys.maxsize,
         labels: set[DocItemLabel] = DEFAULT_EXPORT_LABELS,
         image_mode: ImageRefMode = ImageRefMode.PLACEHOLDER,
+        formula_to_mathml: bool = False,
         page_no: Optional[int] = None,
         html_lang: str = "en",
         html_head: str = _HTML_DEFAULT_HEAD,
@@ -2304,6 +2305,7 @@ def save_as_html(
             to_element=to_element,
             labels=labels,
             image_mode=image_mode,
+            formula_to_mathml=formula_to_mathml,
             page_no=page_no,
             html_lang=html_lang,
             html_head=html_head,

From 16bc35f76918e2216330b742cc59b92402c414e1 Mon Sep 17 00:00:00 2001
From: Michele Dolfi <dol@zurich.ibm.com>
Date: Fri, 31 Jan 2025 13:30:28 +0100
Subject: [PATCH 05/11] rename sanitize in prepare and add \n

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
---
 docling_core/types/doc/document.py | 35 ++++++++++++++++++------------
 1 file changed, 21 insertions(+), 14 deletions(-)

diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
index 83df77bf..ed7c00c4 100644
--- a/docling_core/types/doc/document.py
+++ b/docling_core/types/doc/document.py
@@ -2387,9 +2387,13 @@ def close_lists(
 
         in_ordered_list: List[bool] = []  # False
 
-        def _sanitize_text(text: str, do_escape_html=True) -> str:
+        def _prepare_text(
+            text: str, do_escape_html=True, do_replace_newline=True
+        ) -> str:
             if do_escape_html:
                 text = html.escape(text, quote=False)
+            if do_replace_newline:
+                text = text.replace("\n", "<br>")
             return text
 
         for ix, (item, curr_level) in enumerate(
@@ -2442,7 +2446,7 @@ def _sanitize_text(text: str, do_escape_html=True) -> str:
 
             elif isinstance(item, TextItem) and item.label in [DocItemLabel.TITLE]:
 
-                text = f"<h1>{_sanitize_text(item.text)}</h1>"
+                text = f"<h1>{_prepare_text(item.text)}</h1>"
                 html_texts.append(text.strip())
 
             elif isinstance(item, SectionHeaderItem):
@@ -2451,7 +2455,7 @@ def _sanitize_text(text: str, do_escape_html=True) -> str:
 
                 text = (
                     f"<h{(section_level)}>"
-                    f"{_sanitize_text(item.text)}</h{(section_level)}>"
+                    f"{_prepare_text(item.text)}</h{(section_level)}>"
                 )
                 html_texts.append(text.strip())
 
@@ -2468,16 +2472,18 @@ def _sanitize_text(text: str, do_escape_html=True) -> str:
                     section_level = 6
 
                 text = (
-                    f"<h{section_level}>{_sanitize_text(item.text)}</h{section_level}>"
+                    f"<h{section_level}>{_prepare_text(item.text)}</h{section_level}>"
                 )
                 html_texts.append(text.strip())
 
             elif isinstance(item, TextItem) and item.label in [DocItemLabel.FORMULA]:
 
+                math_formula = _prepare_text(
+                    item.text, do_escape_html=False, do_replace_newline=False
+                )
                 if formula_to_mathml:
                     # Building a math equation in MathML format
                     # ref https://www.w3.org/TR/wai-aria-1.1/#math
-                    math_formula = _sanitize_text(item.text, do_escape_html=False)
                     mathml_element = latex2mathml.converter.convert_to_element(
                         math_formula, display="block"
                     )
@@ -2488,32 +2494,34 @@ def _sanitize_text(text: str, do_escape_html=True) -> str:
                     mathml = unescape(tostring(mathml_element, encoding="unicode"))
                     text = f"<div>{mathml}</div>"
                 else:
-                    text = (
-                        f"<pre>{_sanitize_text(item.text, do_escape_html=False)}</pre>"
-                    )
+                    text = f"<pre>{math_formula}</pre>"
                 html_texts.append(text)
 
             elif isinstance(item, ListItem):
 
-                text = f"<li>{_sanitize_text(item.text)}</li>"
+                text = f"<li>{_prepare_text(item.text)}</li>"
                 html_texts.append(text)
 
             elif isinstance(item, TextItem) and item.label in [DocItemLabel.LIST_ITEM]:
 
-                text = f"<li>{_sanitize_text(item.text)}</li>"
+                text = f"<li>{_prepare_text(item.text)}</li>"
                 html_texts.append(text)
 
             elif isinstance(item, CodeItem):
                 text = (
                     "<pre><code>"
-                    f"{_sanitize_text(item.text, do_escape_html=False)}"
+                    f"{_prepare_text(
+                        item.text,
+                        do_escape_html=False,
+                        do_replace_newline=False
+                    )}"
                     "</code></pre>"
                 )
                 html_texts.append(text.strip())
 
             elif isinstance(item, TextItem):
 
-                text = f"<p>{_sanitize_text(item.text)}</p>"
+                text = f"<p>{_prepare_text(item.text)}</p>"
                 html_texts.append(text.strip())
             elif isinstance(item, TableItem):
 
@@ -2535,8 +2543,7 @@ def _sanitize_text(text: str, do_escape_html=True) -> str:
 
         lines = []
         lines.extend(head_lines)
-        for i, line in enumerate(html_texts):
-            lines.append(line.replace("\n", "<br>"))
+        lines.extend(html_texts)
 
         delim = "\n"
         html_text = (delim.join(lines)).strip()

From dd70bc8c477dff650d3620fb223e3ba400cdaabf Mon Sep 17 00:00:00 2001
From: Michele Dolfi <dol@zurich.ibm.com>
Date: Fri, 31 Jan 2025 13:35:44 +0100
Subject: [PATCH 06/11] fix mypy parsing

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
---
 docling_core/types/doc/document.py | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)

diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
index ed7c00c4..bd5fac2f 100644
--- a/docling_core/types/doc/document.py
+++ b/docling_core/types/doc/document.py
@@ -2508,15 +2508,10 @@ def _prepare_text(
                 html_texts.append(text)
 
             elif isinstance(item, CodeItem):
-                text = (
-                    "<pre><code>"
-                    f"{_prepare_text(
-                        item.text,
-                        do_escape_html=False,
-                        do_replace_newline=False
-                    )}"
-                    "</code></pre>"
+                code_text = _prepare_text(
+                    item.text, do_escape_html=False, do_replace_newline=False
                 )
+                text = f"<pre><code>{code_text}</code></pre>"
                 html_texts.append(text.strip())
 
             elif isinstance(item, TextItem):

From 01383951215f358cb9cb71064137c91e6d3e364e Mon Sep 17 00:00:00 2001
From: Michele Dolfi <dol@zurich.ibm.com>
Date: Fri, 31 Jan 2025 14:08:27 +0100
Subject: [PATCH 07/11] remove unused/impossible elif

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
---
 docling_core/types/doc/document.py | 35 ++++++++----------------------
 1 file changed, 9 insertions(+), 26 deletions(-)

diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
index bd5fac2f..25c4a3fd 100644
--- a/docling_core/types/doc/document.py
+++ b/docling_core/types/doc/document.py
@@ -2387,7 +2387,7 @@ def close_lists(
 
         in_ordered_list: List[bool] = []  # False
 
-        def _prepare_text(
+        def _prepare_tag_content(
             text: str, do_escape_html=True, do_replace_newline=True
         ) -> str:
             if do_escape_html:
@@ -2446,39 +2446,22 @@ def _prepare_text(
 
             elif isinstance(item, TextItem) and item.label in [DocItemLabel.TITLE]:
 
-                text = f"<h1>{_prepare_text(item.text)}</h1>"
+                text = f"<h1>{_prepare_tag_content(item.text)}</h1>"
                 html_texts.append(text.strip())
 
             elif isinstance(item, SectionHeaderItem):
 
-                section_level: int = item.level + 1
+                section_level: int = min(item.level + 1, 6)
 
                 text = (
                     f"<h{(section_level)}>"
-                    f"{_prepare_text(item.text)}</h{(section_level)}>"
-                )
-                html_texts.append(text.strip())
-
-            elif isinstance(item, TextItem) and item.label in [
-                DocItemLabel.SECTION_HEADER
-            ]:
-
-                section_level = curr_level
-
-                if section_level <= 1:
-                    section_level = 2
-
-                if section_level >= 6:
-                    section_level = 6
-
-                text = (
-                    f"<h{section_level}>{_prepare_text(item.text)}</h{section_level}>"
+                    f"{_prepare_tag_content(item.text)}</h{(section_level)}>"
                 )
                 html_texts.append(text.strip())
 
             elif isinstance(item, TextItem) and item.label in [DocItemLabel.FORMULA]:
 
-                math_formula = _prepare_text(
+                math_formula = _prepare_tag_content(
                     item.text, do_escape_html=False, do_replace_newline=False
                 )
                 if formula_to_mathml:
@@ -2499,16 +2482,16 @@ def _prepare_text(
 
             elif isinstance(item, ListItem):
 
-                text = f"<li>{_prepare_text(item.text)}</li>"
+                text = f"<li>{_prepare_tag_content(item.text)}</li>"
                 html_texts.append(text)
 
             elif isinstance(item, TextItem) and item.label in [DocItemLabel.LIST_ITEM]:
 
-                text = f"<li>{_prepare_text(item.text)}</li>"
+                text = f"<li>{_prepare_tag_content(item.text)}</li>"
                 html_texts.append(text)
 
             elif isinstance(item, CodeItem):
-                code_text = _prepare_text(
+                code_text = _prepare_tag_content(
                     item.text, do_escape_html=False, do_replace_newline=False
                 )
                 text = f"<pre><code>{code_text}</code></pre>"
@@ -2516,7 +2499,7 @@ def _prepare_text(
 
             elif isinstance(item, TextItem):
 
-                text = f"<p>{_prepare_text(item.text)}</p>"
+                text = f"<p>{_prepare_tag_content(item.text)}</p>"
                 html_texts.append(text.strip())
             elif isinstance(item, TableItem):
 

From 9dbbfa8fb18f5119a6f387446c942b335cd39287 Mon Sep 17 00:00:00 2001
From: Michele Dolfi <dol@zurich.ibm.com>
Date: Fri, 31 Jan 2025 14:14:22 +0100
Subject: [PATCH 08/11] remove strip()

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
---
 docling_core/types/doc/document.py | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
index 25c4a3fd..299d786f 100644
--- a/docling_core/types/doc/document.py
+++ b/docling_core/types/doc/document.py
@@ -2426,7 +2426,7 @@ def _prepare_tag_content(
             ]:
 
                 text = "<ol>"
-                html_texts.append(text.strip())
+                html_texts.append(text)
 
                 # Increment list nesting level when entering a new list
                 in_ordered_list.append(True)
@@ -2436,7 +2436,7 @@ def _prepare_tag_content(
             ]:
 
                 text = "<ul>"
-                html_texts.append(text.strip())
+                html_texts.append(text)
 
                 # Increment list nesting level when entering a new list
                 in_ordered_list.append(False)
@@ -2447,7 +2447,7 @@ def _prepare_tag_content(
             elif isinstance(item, TextItem) and item.label in [DocItemLabel.TITLE]:
 
                 text = f"<h1>{_prepare_tag_content(item.text)}</h1>"
-                html_texts.append(text.strip())
+                html_texts.append(text)
 
             elif isinstance(item, SectionHeaderItem):
 
@@ -2457,7 +2457,7 @@ def _prepare_tag_content(
                     f"<h{(section_level)}>"
                     f"{_prepare_tag_content(item.text)}</h{(section_level)}>"
                 )
-                html_texts.append(text.strip())
+                html_texts.append(text)
 
             elif isinstance(item, TextItem) and item.label in [DocItemLabel.FORMULA]:
 
@@ -2495,12 +2495,12 @@ def _prepare_tag_content(
                     item.text, do_escape_html=False, do_replace_newline=False
                 )
                 text = f"<pre><code>{code_text}</code></pre>"
-                html_texts.append(text.strip())
+                html_texts.append(text)
 
             elif isinstance(item, TextItem):
 
                 text = f"<p>{_prepare_tag_content(item.text)}</p>"
-                html_texts.append(text.strip())
+                html_texts.append(text)
             elif isinstance(item, TableItem):
 
                 text = item.export_to_html(doc=self, add_caption=True)

From eb0cac44f6725c6d3d025319bae6271d78132853 Mon Sep 17 00:00:00 2001
From: Michele Dolfi <dol@zurich.ibm.com>
Date: Fri, 31 Jan 2025 14:17:05 +0100
Subject: [PATCH 09/11] add display none for latex annotation

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
---
 docling_core/types/doc/document.py                | 3 +++
 test/data/doc/2206.01062.yaml.html                | 3 +++
 test/data/doc/bad_doc.yaml.html                   | 3 +++
 test/data/doc/constructed_doc.embedded.html.gt    | 3 +++
 test/data/doc/constructed_doc.placeholder.html.gt | 3 +++
 test/data/doc/constructed_doc.referenced.html.gt  | 3 +++
 test/data/doc/constructed_document.yaml.html      | 3 +++
 test/data/doc/dummy_doc.yaml.html                 | 3 +++
 8 files changed, 24 insertions(+)

diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
index 299d786f..eb56e7cc 100644
--- a/docling_core/types/doc/document.py
+++ b/docling_core/types/doc/document.py
@@ -1390,6 +1390,9 @@ class DoclingDocument(BaseModel):
     table tr:nth-child(even) td{
     background-color: LightGray;
     }
+    math annotation {
+    display: none;
+    }
     </style>
     </head>"""
 
diff --git a/test/data/doc/2206.01062.yaml.html b/test/data/doc/2206.01062.yaml.html
index ba2add20..ea69871b 100644
--- a/test/data/doc/2206.01062.yaml.html
+++ b/test/data/doc/2206.01062.yaml.html
@@ -53,6 +53,9 @@
     table tr:nth-child(even) td{
     background-color: LightGray;
     }
+    math annotation {
+    display: none;
+    }
     </style>
     </head>
 <h2>DocLayNet: A Large Human-Annotated Dataset for Document-Layout Analysis</h2>
diff --git a/test/data/doc/bad_doc.yaml.html b/test/data/doc/bad_doc.yaml.html
index 12cf61d0..c6c20f87 100644
--- a/test/data/doc/bad_doc.yaml.html
+++ b/test/data/doc/bad_doc.yaml.html
@@ -53,6 +53,9 @@
     table tr:nth-child(even) td{
     background-color: LightGray;
     }
+    math annotation {
+    display: none;
+    }
     </style>
     </head>
 <h1>This is the title</h1>
diff --git a/test/data/doc/constructed_doc.embedded.html.gt b/test/data/doc/constructed_doc.embedded.html.gt
index 12ef29ee..ec63d1a2 100644
--- a/test/data/doc/constructed_doc.embedded.html.gt
+++ b/test/data/doc/constructed_doc.embedded.html.gt
@@ -53,6 +53,9 @@
     table tr:nth-child(even) td{
     background-color: LightGray;
     }
+    math annotation {
+    display: none;
+    }
     </style>
     </head>
 <h1>Title of the Document</h1>
diff --git a/test/data/doc/constructed_doc.placeholder.html.gt b/test/data/doc/constructed_doc.placeholder.html.gt
index e63df310..f14a80e0 100644
--- a/test/data/doc/constructed_doc.placeholder.html.gt
+++ b/test/data/doc/constructed_doc.placeholder.html.gt
@@ -53,6 +53,9 @@
     table tr:nth-child(even) td{
     background-color: LightGray;
     }
+    math annotation {
+    display: none;
+    }
     </style>
     </head>
 <h1>Title of the Document</h1>
diff --git a/test/data/doc/constructed_doc.referenced.html.gt b/test/data/doc/constructed_doc.referenced.html.gt
index da449dfb..4678f9cb 100644
--- a/test/data/doc/constructed_doc.referenced.html.gt
+++ b/test/data/doc/constructed_doc.referenced.html.gt
@@ -53,6 +53,9 @@
     table tr:nth-child(even) td{
     background-color: LightGray;
     }
+    math annotation {
+    display: none;
+    }
     </style>
     </head>
 <h1>Title of the Document</h1>
diff --git a/test/data/doc/constructed_document.yaml.html b/test/data/doc/constructed_document.yaml.html
index e63df310..f14a80e0 100644
--- a/test/data/doc/constructed_document.yaml.html
+++ b/test/data/doc/constructed_document.yaml.html
@@ -53,6 +53,9 @@
     table tr:nth-child(even) td{
     background-color: LightGray;
     }
+    math annotation {
+    display: none;
+    }
     </style>
     </head>
 <h1>Title of the Document</h1>
diff --git a/test/data/doc/dummy_doc.yaml.html b/test/data/doc/dummy_doc.yaml.html
index 7de2efc7..46d5c090 100644
--- a/test/data/doc/dummy_doc.yaml.html
+++ b/test/data/doc/dummy_doc.yaml.html
@@ -53,6 +53,9 @@
     table tr:nth-child(even) td{
     background-color: LightGray;
     }
+    math annotation {
+    display: none;
+    }
     </style>
     </head>
 <h1>DocLayNet: A Large Human-Annotated Dataset for Document-Layout Analysis</h1>

From 25c681328268af038e6246b533c5fba16fdd4c3e Mon Sep 17 00:00:00 2001
From: Michele Dolfi <dol@zurich.ibm.com>
Date: Fri, 31 Jan 2025 15:24:27 +0100
Subject: [PATCH 10/11] make mathml the default

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
---
 docling_core/types/doc/document.py | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
index eb56e7cc..a3616a4a 100644
--- a/docling_core/types/doc/document.py
+++ b/docling_core/types/doc/document.py
@@ -2288,7 +2288,7 @@ def save_as_html(
         to_element: int = sys.maxsize,
         labels: set[DocItemLabel] = DEFAULT_EXPORT_LABELS,
         image_mode: ImageRefMode = ImageRefMode.PLACEHOLDER,
-        formula_to_mathml: bool = False,
+        formula_to_mathml: bool = True,
         page_no: Optional[int] = None,
         html_lang: str = "en",
         html_head: str = _HTML_DEFAULT_HEAD,
@@ -2355,7 +2355,7 @@ def export_to_html(  # noqa: C901
         to_element: int = sys.maxsize,
         labels: set[DocItemLabel] = DEFAULT_EXPORT_LABELS,
         image_mode: ImageRefMode = ImageRefMode.PLACEHOLDER,
-        formula_to_mathml: bool = False,
+        formula_to_mathml: bool = True,
         page_no: Optional[int] = None,
         html_lang: str = "en",
         html_head: str = _HTML_DEFAULT_HEAD,
@@ -2479,7 +2479,13 @@ def _prepare_tag_content(
                     annotation.text = math_formula
                     mathml = unescape(tostring(mathml_element, encoding="unicode"))
                     text = f"<div>{mathml}</div>"
-                else:
+
+                elif (
+                    item.text == ""
+                    and item.orig != ""
+                    and image_mode == ImageRefMode.EMBEDDED
+                ):
+
                     text = f"<pre>{math_formula}</pre>"
                 html_texts.append(text)
 

From 4af63cdc4ff100b136b06cba327294411765e4f3 Mon Sep 17 00:00:00 2001
From: Michele Dolfi <dol@zurich.ibm.com>
Date: Fri, 31 Jan 2025 15:26:11 +0100
Subject: [PATCH 11/11] revert wrong commit

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
---
 docling_core/types/doc/document.py | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
index a3616a4a..0f463001 100644
--- a/docling_core/types/doc/document.py
+++ b/docling_core/types/doc/document.py
@@ -2480,12 +2480,7 @@ def _prepare_tag_content(
                     mathml = unescape(tostring(mathml_element, encoding="unicode"))
                     text = f"<div>{mathml}</div>"
 
-                elif (
-                    item.text == ""
-                    and item.orig != ""
-                    and image_mode == ImageRefMode.EMBEDDED
-                ):
-
+                else:
                     text = f"<pre>{math_formula}</pre>"
                 html_texts.append(text)