fix: preserve XLSX currency formatting

packages/markitdown/src/markitdown/converters/_xlsx_converter.py

@@ -1,4 +1,6 @@
 import sys
+import re
+import numbers
 from typing import BinaryIO, Any
 from ._html_converter import HtmlConverter
 from .._base_converter import DocumentConverter, DocumentConverterResult
@@ -10,7 +12,7 @@
 _xlsx_dependency_exc_info = None
 try:
     import pandas as pd
-    import openpyxl  # noqa: F401
+    import openpyxl
 except ImportError:
     _xlsx_dependency_exc_info = sys.exc_info()
 
@@ -32,6 +34,30 @@
 ]
 ACCEPTED_XLS_FILE_EXTENSIONS = [".xls"]
 
+_CURRENCY_SYMBOL_RE = re.compile(
+    r"(?<!\\)([$\u20ac\u00a3\u00a5])|\[\$([^-\]]+)(?:-[^\]]+)?\]"
+)
+
+
+def _format_currency_value(value: Any, number_format: str) -> str | None:
+    match = _CURRENCY_SYMBOL_RE.search(number_format)
+    if (
+        match is None
+        or isinstance(value, bool)
+        or not isinstance(value, numbers.Real)
+        or pd.isna(value)
+    ):
+        return None
+
+    symbol = match.group(1) or match.group(2)
+    positive_format = number_format.split(";", 1)[0]
+    decimal_part = positive_format.split(".", 1)[1] if "." in positive_format else ""
+    precision = sum(1 for char in decimal_part if char in "0#")
+    use_grouping = "," in positive_format.split(".", 1)[0]
+
+    grouping = "," if use_grouping else ""
+    return f"{symbol}{value:{grouping}.{precision}f}"
+
 
 class XlsxConverter(DocumentConverter):
     """
@@ -81,8 +107,25 @@ def convert(
             )
 
         sheets = pd.read_excel(file_stream, sheet_name=None, engine="openpyxl")
+        file_stream.seek(0)
+        workbook = openpyxl.load_workbook(file_stream, read_only=True, data_only=True)
+
         md_content = ""
         for s in sheets:
+            worksheet = workbook[s]
+            for row_idx in range(len(sheets[s].index)):
+                for col_idx in range(len(sheets[s].columns)):
+                    cell = worksheet.cell(row=row_idx + 2, column=col_idx + 1)
+                    currency_value = _format_currency_value(
+                        sheets[s].iat[row_idx, col_idx],
+                        cell.number_format,
+                    )
+                    if currency_value is not None:
+                        sheets[s][sheets[s].columns[col_idx]] = sheets[s][
+                            sheets[s].columns[col_idx]
+                        ].astype(object)
+                        sheets[s].iat[row_idx, col_idx] = currency_value
+
             md_content += f"## {s}\n"
             html_content = sheets[s].to_html(index=False)
             md_content += (

packages/markitdown/tests/test_module_misc.py

@@ -288,6 +288,24 @@ def test_input_as_strings() -> None:
     assert "# Test" in result.text_content
 
 
+def test_xlsx_currency_format_is_preserved(tmp_path) -> None:
+    from openpyxl import Workbook
+
+    workbook = Workbook()
+    worksheet = workbook.active
+    worksheet.title = "Invoice"
+    worksheet.append(["Item", "Count", "Cost"])
+    worksheet.append(["Breakfast", 20, 5])
+    worksheet["C2"].number_format = "$#,##0.00"
+
+    xlsx_path = tmp_path / "currency.xlsx"
+    workbook.save(xlsx_path)
+
+    result = MarkItDown().convert(str(xlsx_path))
+
+    assert "| Breakfast | 20 | $5.00 |" in result.markdown
+
+
 def test_deeply_nested_html_fallback() -> None:
     """Large, deeply nested HTML should fall back to plain-text extraction
     instead of silently returning unconverted HTML (issue #1636).