[bug] Strikethrough-only cells are treated as unstyled

[arnav2]

Summary

Cells whose only non-default font attribute is strike=True lose their style entirely: cell.style comes back as None. The same happens for underline-only cells.

Reproduction

from pathlib import Path
from tempfile import TemporaryDirectory

import openpyxl
from openpyxl.styles import Font
from ks_xlsx_parser import parse_workbook

with TemporaryDirectory() as td:
    path = Path(td) / "strike_test.xlsx"
    wb = openpyxl.Workbook()
    ws = wb.active
    ws["A1"] = "Normal"
    ws["B1"] = "Strike"
    ws["B1"].font = Font(strike=True)
    ws["C1"] = "BoldStrike"
    ws["C1"].font = Font(bold=True, strike=True, color="FF0000")
    wb.save(path)

    sheet = parse_workbook(path=path).workbook.sheets[0]
    for col in range(1, 4):
        cell = sheet.get_cell(1, col)
        font = cell.style.font if cell.style and cell.style.font else None
        print(cell.coord.to_a1(), "has_style=", bool(cell.style), "font=",
              font.model_dump(exclude_none=True) if font else None)

Actual

A1 has_style= True  font= {'name': 'Calibri', 'size': 11.0, ..., 'strikethrough': False, 'color': 'theme:1'}
B1 has_style= False font= None
C1 has_style= True  font= {'bold': True, 'strikethrough': True, 'color': 'FF0000'}

Expected: B1 keeps its style and reports strikethrough: true.

C1 works only incidentally — bold/color are what keep the style alive there, not the strike.

Cause

_extract_font() in src/parsers/cell_parser.py parses strikethrough correctly, but _extract_style() decides whether a cell is styled from a hand-written attribute list that omits both strikethrough and underline:

has_style = any([
    font and (font.bold or font.italic or font.name or font.size or font.color),
    fill and fill.fg_color,
    border and any([border.left, border.right, border.top, border.bottom]),
    alignment and (alignment.horizontal or alignment.vertical or alignment.wrap_text),
    number_format,
])

When the whole CellStyle is discarded, the parsed font (strike included) goes with it.

The enumeration also under-reports two other cases: a fill with only bg_color, and an alignment with only text_rotation/indent. Each _extract_* helper already returns None when nothing is set, so the enumeration is redundant as well as lossy — comparing against a default DTO covers all of them:

has_style = any([
    font is not None and font != _DEFAULT_FONT,
    fill is not None,
    border is not None,
    alignment is not None,
    number_format,
])

Secondary: expose strikethrough in to_json()

Chunk cells in result.to_json()["chunks"][...]["cells"] currently carry address, value, formula, font_color, fill_color. Strikethrough is widely used in real workbooks to mark deprecated or void rows, so downstream RAG/LLM consumers need it at that level too, alongside the colors:

{"address": "B1", "value": "Strike", "formula": null,
 "font_color": null, "fill_color": null, "font_strikethrough": true}

Environment

Reported against ks-xlsx-parser 0.2.1, Python 3.13, Windows. Reproduced on main.

Reported by a user via project feedback.