excel-model 0.1.5

YAML-driven Excel financial model generator

excel-model

YAML-driven Excel financial model generator.

Build professional financial models (P&L, DCF, Budget vs Actuals, Scenario Analysis) from declarative YAML specs. Generates .xlsx workbooks with named ranges, styled sheets, and Excel formulas using openpyxl.

Documentation

Installation

pip install excel-model

Or for development:

pixi install

Quick Start

CLI

# Build a P&L model
excel-model build --spec model.yaml --output model.xlsx --mode batch

# Validate a spec
excel-model validate --spec model.yaml

# Describe what a spec would produce
excel-model describe --spec model.yaml --format text

Python API

from excel_model.spec_loader import load_spec
from excel_model.validator import validate_spec
from excel_model.excel_writer import build_workbook
from excel_model.config import load_style

spec = load_spec("model.yaml")
errors = validate_spec(spec)
assert not errors

style = load_style(None)  # uses bundled defaults
build_workbook(spec=spec, inputs=None, output_path="model.xlsx", style=style)

Standalone Data Sheets

from openpyxl import Workbook
from excel_model.config import load_style
from excel_model import DataSheetDef, SumifsPivotDef, build_data_sheet, build_sumifs_pivot

wb = Workbook()
del wb["Sheet"]
style = load_style(None)

# Simple tabular data sheet
data_spec = DataSheetDef(
    sheet_name="DATA", title="My Data",
    headers=("Name", "Amount"),
    col_widths=(20.0, 14.0), number_formats={1: "#,##0"}, freeze_row=2,
)
build_data_sheet(wb=wb, spec=data_spec, rows=[["Vendor A", 12345]], style=style)

# SUMIFS pivot sheet (formulas reference a source data sheet)
pivot_spec = SumifsPivotDef(
    sheet_name="PIVOT", title="Revenue by Hub",
    row_label_headers=("Hub",), col_dim_values=(2023, 2024, 2025),
    data_sheet="DATA", value_col="B",
    row_filter_cols=("A",), col_filter_col="C",
    append_total=True, append_yoy=True,
    col_widths=(28.0, 14.0, 14.0, 14.0, 16.0, 12.0, 12.0),
    number_format_data="#,##0", number_format_pct="0.0%", freeze_row=2,
)
build_sumifs_pivot(wb=wb, spec=pivot_spec, row_labels=[["Berlin"], ["Munich"]], style=style)

wb.save("output.xlsx")

Model Types

Type	Description
p_and_l	Profit & Loss statement
dcf	Discounted Cash Flow valuation
budget_vs_actuals	Budget vs Actuals comparison
scenario	Multi-scenario analysis (Base/Bull/Bear)
comparison	Cross-entity comparison

Formula Types

21 built-in formula types including growth_projected, pct_of_revenue, sum_of_rows, subtraction, ratio, discounted_pv, terminal_value, npv_sum, variance, variance_pct, constant, custom, and more.

Custom formula security: The custom formula type rejects formulas containing dangerous patterns (DDE, WEBSERVICE, IMPORTDATA, CALL, EXEC, FILTERXML, REGISTER.ID, etc.) to prevent Excel formula injection attacks. Standard Excel functions like SUM, IF, ROUND, and MAX are allowed.

Configuration

Style config controls Excel formatting (colors, fonts, number formats). A bundled default is included; override with --style:

header_fill_hex: "1F3864"
header_font_color: "FFFFFF"
font_name: "Calibri"
font_size: 10
number_format_currency: '#,##0'
number_format_percent: '0.0%'

Security Note

File path arguments (--spec, --data, --style, --output) are passed directly to the filesystem without path containment checks. This is safe for the default CLI usage where the authenticated user controls their own filesystem. If you wrap this tool in a web API or automated pipeline that accepts user-controlled path inputs, you must validate that resolved paths stay within an allowed base directory before invoking the CLI to prevent path traversal attacks.

Looking for Financial Modeling Input

This library was built by a software engineer, not a financial analyst. The model structures, formula types, and default assumptions reflect a developer's interpretation of common financial models.

If you work in finance, FP&A, investment banking, or accounting, your input would be incredibly valuable:

• Are the formula types correct? Do growth_projected, pct_of_revenue, discounted_pv, and terminal_value follow standard conventions?
• Missing model patterns? Are there common financial model structures (e.g., waterfall, three-statement, LBO) that should be supported?
• Named range conventions -- do the Excel named range naming patterns match what analysts expect?
• Number formatting -- are the default currency/percent/integer formats appropriate for professional models?
• Scenario analysis -- does the base/bull/bear override pattern match how scenarios are typically structured?

Please open an issue with the type:feat label, or start a discussion. All feedback is welcome, from quick corrections to detailed model reviews.

License

MIT