sheetpeek

sheetpeek shows an Excel workbook the way Excel shows it: as HTML, Markdown, terminal output or plain text. It is pure Python, needs no office suite, and gives programs access to the rendered cells. Number formats are applied by xlformat, whose output is tested against text captured from Excel 16 for Mac.

Documentation: https://sheetpeek.readthedocs.io.

Reading a sheet

read returns a Sheet: the cached values, number formats, styles, column widths, row heights and merged ranges of one worksheet. Dates stay the serial numbers Excel stores.

>>> import tempfile, pathlib, openpyxl, sheetpeek
>>> path = pathlib.Path(tempfile.mkdtemp()) / "example.xlsx"
>>> wb = openpyxl.Workbook()
>>> wb.active["A1"] = 1234.5
>>> wb.active["A1"].number_format = "#,##0.00"
>>> wb.active.merge_cells("B1:C1")
>>> wb.save(path)
>>> sheet = sheetpeek.read(path)
>>> cell = sheet.cell(1, 1)
>>> cell.value, cell.kind, cell.number_format
(1234.5, 'number', '#,##0.00')
>>> [str(m) for m in sheet.merges]
['B1:C1']

Showing a sheet

sheetpeek.render draws a sheet with a backend. Each shows what Excel shows, as far as its format can: number formats, # for a number too wide for its column, text spilling into empty neighbors.

>>> path = pathlib.Path(tempfile.mkdtemp()) / "sales.xlsx"
>>> wb = openpyxl.Workbook()
>>> ws = wb.active
>>> ws.append(["Region", "Sales", "Change"])
>>> ws.append(["North", 1234.5, 0.125])
>>> ws.append(["South", -98765.4, -0.0425])
>>> ws.append(["A note long enough to spill"])
>>> for row in (2, 3):
...     ws.cell(row, 2).number_format = "#,##0.00;[Red]-#,##0.00"
...     ws.cell(row, 3).number_format = "0.0%"
>>> wb.save(path)
>>> sales = sheetpeek.read(path)
>>> print(sheetpeek.render(sales, "text"), end="")
Region   Sales    Change
North    1,234.50    12.5%
South    ########    -4.3%
A note long enough to spill
>>> print(sheetpeek.render(sales, "markdown"), end="")
| Region | Sales | Change |
| --- | ---: | ---: |
| North | 1,234.50 | 12.5% |
| South | -98,765.40 | -4.3% |
| A note long enough to spill |  |  |
>>> print(sheetpeek.render(sales, "csv"), end="")
Region,Sales,Change
North,"1,234.50",12.5%
South,"-98,765.40",-4.3%
A note long enough to spill,,
>>> page = sheetpeek.render(sales, "html")  # a static page in Excel's own geometry
>>> page.startswith("<!doctype html>")
True

Fixed-width text gives every column its width in characters; Markdown and CSV have no widths, so they show every value in full. HTML is drawn in points from Excel's layout: column widths, row heights, each text's position and clip window, the # runs in the section's color (the negative is red), and fills. sheetpeek.render(sales, "rich"), or rich.print(sales), draws the fixed-width layout in a terminal with fonts, colors and fills (the optional rich extra).

From the command line:

sheetpeek view sales.xlsx                 # rich on a terminal, plain text when piped
sheetpeek view sales.xlsx -t html > sales.html
sheetpeek serve reports/                  # every workbook under reports/ as HTML pages

Sheet.layout gives the layout itself, the data every backend draws from: in excel mode each column's position and width and each text's position in points, in characters mode the characters each column draws. The documentation's usage page has examples.

Cells are read with openpyxl's read-only mode, so memory does not grow with the file while it is read. The parts read-only mode leaves out, such as merges, widths and the used range, come from a separate scan of the sheet's XML. ref="A1:H40" reads part of a sheet, and sheetpeek.open reads several sheets with one load.

Installation

pip install sheetpeek
pip install "sheetpeek[rich]"   # with styled terminal output

Python 3.11 or later. Depends on openpyxl and xlformat.

Related packages

sheetpeek is part of the sheetmill project, with xlformat, which applies Excel number formats. A third package, a formula engine, is in development.

Development

uv sync --group docs then uv run pytest; CONTRIBUTING.md has the pip route, the full set of checks and the commit conventions. ARCHITECTURE.md describes the design and the measurements behind it.

Most of the code is written with Claude Code, reviewed before each commit and tested against output captured from Excel; see CONTRIBUTING.md.

License

MIT.