feat: preserve Excel cell formatting via XlsxConfig (fix #53)

[vanity (Vanity-1)]

Fixes #53

Summary

Adds an opt-in XlsxConfig object to preserve Excel cell number formatting (currency, percentage, thousands separators, dates) when converting .xlsx files to Markdown.

Motivation

pd.read_excel only reads raw cell values, so formatted values like $1,199.00 are emitted as 1199. This PR makes formatting preservation available without breaking existing consumers that parse numeric values from the output.

Changes

• Add XlsxConfig dataclass with a preserve_formatting: bool = False flag (default keeps current behavior).
• XlsxConverter now accepts an optional config argument; when preserve_formatting is enabled it reads cells via openpyxl and renders values according to their number format.
• Currency symbols are extracted from the format string (e.g. $, €, ¥, []) rather than hardcoded.
• MarkItDown(xlsx_config=XlsxConfig(preserve_formatting=True)) wires the config through �nable_builtins.
• XlsxConfig is exported from the markitdown package.
• Adds ests/test_xlsx_converter.py covering currency, percentage, thousands separators, negative values, other currency symbols, dates, and backward compatibility.

Usage

python
from markitdown import MarkItDown, XlsxConfig

md = MarkItDown(xlsx_config=XlsxConfig(preserve_formatting=True))
result = md.convert("invoice.xlsx")

Backward compatibility

Default behavior is unchanged (raw cell values), so existing code that does loat(cell_value) keeps working.

[vanity (Vanity-1)]

@microsoft-github-policy-service agree