Skip to content

RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.

fix(parser): escape DOCX table cells in the naive, manual and qa chunkers (#19746)

Lukasba693dac0346
diff --git a/rag/app/manual.py b/rag/app/manual.py
index 80f282dedb..165e6b3263 100644
--- a/rag/app/manual.py
+++ b/rag/app/manual.py
@@ -17,6 +17,7 @@
import copy
import logging
import re
+from html import escape as html_escape
from io import BytesIO
from docx import Document
@@ -134,7 +135,8 @@ class Docx(DocxParser):
else:
break
i += 1
- html += f"<td>{c.text}</td>" if span == 1 else f"<td colspan='{span}'>{c.text}</td>"
+ cell = html_escape(c.text)
+ html += f"<td>{cell}</td>" if span == 1 else f"<td colspan='{span}'>{cell}</td>"
html += "</tr>"
html += "</table>"
tbls.append(((None, html), ""))
diff --git a/rag/app/naive.py b/rag/app/naive.py
index 672e031ef3..02bd468dfd 100644
--- a/rag/app/naive.py
+++ b/rag/app/naive.py
@@ -19,6 +19,7 @@ import os
import re
from collections.abc import Callable
from functools import reduce
+from html import escape as html_escape
from io import BytesIO
from timeit import default_timer as timer
from typing import Any
@@ -812,7 +813,7 @@ class Docx(DocxParser):
title = self.__get_nearest_title(table_idx, filename)
html = "<table>"
if title:
- html += f"<caption>Table Location: {title}</caption>"
+ html += f"<caption>Table Location: {html_escape(title)}</caption>"
for r in tb.rows:
html += "<tr>"
col_idx = 0
@@ -827,7 +828,8 @@ class Docx(DocxParser):
else:
break
col_idx += 1
- html += f"<td>{c.text}</td>" if span == 1 else f"<td colspan='{span}'>{c.text}</td>"
+ cell = html_escape(c.text)
+ html += f"<td>{cell}</td>" if span == 1 else f"<td colspan='{span}'>{cell}</td>"
except Exception as e:
logging.warning(f"Error parsing table, ignore: {e}")
html += "</tr>"
diff --git a/rag/app/qa.py b/rag/app/qa.py
index a97b890611..c02f105331 100644
--- a/rag/app/qa.py
+++ b/rag/app/qa.py
@@ -18,6 +18,7 @@ import csv
import logging
import re
from copy import deepcopy
+from html import escape as html_escape
from io import BytesIO
from timeit import default_timer as timer
@@ -240,7 +241,8 @@ class Docx(DocxParser):
else:
break
i += 1
- html += f"<td>{c.text}</td>" if span == 1 else f"<td colspan='{span}'>{c.text}</td>"
+ cell = html_escape(c.text)
+ html += f"<td>{cell}</td>" if span == 1 else f"<td colspan='{span}'>{cell}</td>"
html += "</tr>"
html += "</table>"
tbls.append(((None, html), ""))
diff --git a/test/unit_test/rag/test_docx_table_escaping.py b/test/unit_test/rag/test_docx_table_escaping.py
new file mode 100644
index 0000000000..f285243e1d
--- /dev/null
+++ b/test/unit_test/rag/test_docx_table_escaping.py
@@ -0,0 +1,225 @@
+#
+# Copyright 2025 The InfiniFlow Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+"""Three DOCX chunkers built their table HTML without escaping the cells.
+
+`rag/app/laws.py` escapes them (`test_laws_docx_tables.py` pins it), and so does
+`deepdoc/parser/excel_parser.py`. `naive.py`, `manual.py` and `qa.py` did not, so
+a cell holding `<`, `>` or `&` -- a comparison, a formula, an XML snippet, a
+company name -- was written into the table as markup.
+"""
+
+import importlib.util
+import sys
+import types
+from io import BytesIO
+from pathlib import Path
+from unittest import mock
+
+import pytest
+
+_REPO_ROOT = Path(__file__).resolve().parents[3]
+
+CELL_A = "a < b & c > d"
+CELL_B = "<img src=x onerror=alert(1)>"
+ESCAPED_A = "a &lt; b &amp; c &gt; d"
+ESCAPED_B = "&lt;img src=x onerror=alert(1)&gt;"
+
+
+class _Dummy:
+ def __init__(self, *a, **k):
+ pass
+
+
+def _stub(name, **attrs):
+ mod = types.ModuleType(name)
+ for key, value in attrs.items():
+ setattr(mod, key, value)
+ sys.modules[name] = mod
+ return mod
+
+
+def _load(module_name, relative_path):
+ spec = importlib.util.spec_from_file_location(module_name, _REPO_ROOT / relative_path)
+ module = importlib.util.module_from_spec(spec)
+ sys.modules[module_name] = module
+ spec.loader.exec_module(module)
+ return module
+
+
[email protected](scope="module")
+def chunkers():
+ """The DOCX chunkers, with the vision / LLM / storage siblings stubbed out."""
+ saved = {name: sys.modules.get(name) for name in list(sys.modules)}
+ try:
+ for name in ("xgboost", "pdfplumber", "huggingface_hub", "pypdf", "sklearn", "sklearn.cluster", "sklearn.metrics", "deepdoc.vision"):
+ sys.modules.setdefault(name, mock.MagicMock())
+
+ docx_parser = _load("deepdoc.parser.docx_parser_under_test", "deepdoc/parser/docx_parser.py")
+
+ _stub(
+ "deepdoc.parser",
+ DocxParser=docx_parser.RAGFlowDocxParser,
+ PdfParser=_Dummy,
+ HtmlParser=_Dummy,
+ ExcelParser=_Dummy,
+ EpubParser=_Dummy,
+ JsonParser=_Dummy,
+ MarkdownElementExtractor=_Dummy,
+ MarkdownParser=_Dummy,
+ TxtParser=_Dummy,
+ )
+ _stub("deepdoc.parser.utils", get_text=lambda *a, **k: "", extract_pdf_outlines=lambda *a, **k: [])
+ _stub(
+ "deepdoc.parser.figure_parser",
+ VisionFigureParser=_Dummy,
+ vision_figure_parser_pdf_wrapper=lambda *a, **k: [],
+ vision_figure_parser_docx_wrapper=lambda *a, **k: [],
+ vision_figure_parser_docx_wrapper_naive=lambda *a, **k: [],
+ )
+ _stub("deepdoc.parser.pdf_parser", PlainParser=_Dummy, VisionParser=_Dummy, RAGFlowPdfParser=_Dummy)
+ _stub("deepdoc.parser.docling_parser", DoclingParser=_Dummy)
+ _stub("deepdoc.parser.monkeyocrv2_parser", MonkeyOCRv2Parser=_Dummy)
+ _stub("deepdoc.parser.tcadp_parser", TCADPParser=_Dummy)
+ _stub(
+ "api.db.joint_services.tenant_model_service",
+ get_composite_model_name_by_id=lambda x: x,
+ get_first_provider_model_name=lambda *a, **k: None,
+ resolve_model_config=lambda *a, **k: None,
+ get_tenant_default_model_by_type=lambda *a, **k: None,
+ ensure_mineru_from_env=lambda *a, **k: None,
+ ensure_opendataloader_from_env=lambda *a, **k: None,
+ ensure_paddleocr_from_env=lambda *a, **k: None,
+ ensure_somark_from_env=lambda *a, **k: None,
+ )
+ _stub("api.db.services.llm_service", LLMBundle=_Dummy)
+ _stub(
+ "rag.utils.file_utils",
+ extract_embed_file=lambda *a, **k: None,
+ extract_links_from_pdf=lambda *a, **k: [],
+ extract_links_from_docx=lambda *a, **k: [],
+ extract_html=lambda *a, **k: (None, None),
+ )
+ _stub("common.float_utils", get_float=lambda *a, **k: 0.0, normalize_overlapped_percent=lambda x: x)
+ _stub("common.text_utils", normalize_arabic_presentation_forms=lambda x: x)
+ _stub("rag.app.naive", by_plaintext=lambda *a, **k: ([], [], None), PARSERS={})
+ _stub(
+ "common.parser_config_utils",
+ normalize_layout_recognizer=lambda x: (x, None),
+ has_mineru_options=lambda *a, **k: False,
+ is_tenant_model_id=lambda *a, **k: False,
+ MINERU_OPTION_KEYS=(),
+ )
+
+ yield types.SimpleNamespace(
+ laws=_load("rag.app.laws_under_test", "rag/app/laws.py").Docx,
+ qa=_load("rag.app.qa_under_test", "rag/app/qa.py").Docx,
+ manual=_load("rag.app.manual_under_test", "rag/app/manual.py").Docx,
+ naive=_load("rag.app.naive_under_test", "rag/app/naive.py").Docx,
+ )
+ finally:
+ for name in list(sys.modules):
+ if name not in saved:
+ del sys.modules[name]
+ sys.modules.update(saved)
+
+
+def _docx_with_table(heading="Heading"):
+ from docx import Document
+
+ doc = Document()
+ doc.add_heading(heading, level=1)
+ table = doc.add_table(rows=1, cols=2)
+ table.cell(0, 0).text = CELL_A
+ table.cell(0, 1).text = CELL_B
+
+ buf = BytesIO()
+ doc.save(buf)
+ return buf.getvalue()
+
+
+def _assert_escaped(html):
+ assert ESCAPED_A in html
+ assert ESCAPED_B in html
+ # The markup a cell carried must not reach the table as markup.
+ assert "<img" not in html
+ assert "<td>a < b" not in html
+
+
+def test_naive_docx_escapes_cell_html(chunkers):
+ lines = chunkers.naive()("table.docx", _docx_with_table())
+
+ tables = [table for _text, _image, table in lines if table]
+ assert tables
+ _assert_escaped(tables[0])
+
+
+def test_naive_docx_escapes_the_table_caption(chunkers):
+ """The caption carries the enclosing heading, which is document text too."""
+ lines = chunkers.naive()("table.docx", _docx_with_table(heading="Q1 < Q2 & Q3"))
+
+ tables = [table for _text, _image, table in lines if table]
+ assert tables
+ assert "<caption>" in tables[0]
+ assert "Q1 &lt; Q2 &amp; Q3" in tables[0]
+ assert "Q1 < Q2" not in tables[0]
+
+
+def test_manual_docx_escapes_cell_html(chunkers):
+ _ti_list, tbls = chunkers.manual()("table.docx", _docx_with_table())
+
+ assert tbls
+ _assert_escaped(tbls[0][0][1])
+
+
+def test_qa_docx_escapes_cell_html(chunkers):
+ _qai_list, tbls = chunkers.qa()("table.docx", _docx_with_table())
+
+ assert tbls
+ _assert_escaped(tbls[0][0][1])
+
+
+def test_laws_docx_still_escapes_cell_html(chunkers):
+ """Guard: the chunker that already escaped must not change."""
+ chunks = chunkers.laws()("table.docx", _docx_with_table())
+
+ table_chunk = next(chunk for chunk in chunks if "<table>" in chunk)
+ _assert_escaped(table_chunk)
+
+
+def test_a_plain_table_is_unchanged(chunkers):
+ """Guard: a table with nothing to escape renders exactly as before."""
+ from docx import Document
+
+ doc = Document()
+ doc.add_heading("Heading", level=1)
+ table = doc.add_table(rows=1, cols=2)
+ table.cell(0, 0).text = "Item"
+ table.cell(0, 1).text = "Fee"
+ buf = BytesIO()
+ doc.save(buf)
+
+ lines = chunkers.naive()("table.docx", buf.getvalue())
+ tables = [table for _text, _image, table in lines if table]
+
+ assert "<td>Item</td>" in tables[0]
+ assert "<td>Fee</td>" in tables[0]