Marimo SciPy Utils

Utility functions for creating interactive marimo components with scipy distributions.

This package provides functions for creating and configuring interactive UI elements in marimo notebooks, with a focus on parameter input and visualization using scipy probability distributions.

Installation

uv add marimo-scipy-utils

Functions

See example notebook here

abbrev_format(x: float, pos: int | None) -> str

Format numbers with k/M suffixes for thousands/millions.

Parameters:

• x: The number to format
• pos: The tick position (unused but required by matplotlib formatter interface)

Returns:

• str: The formatted number string with k/M suffix if applicable

Example:

abbrev_format(1500, None)  # Returns "2k"
abbrev_format(2500000, None)  # Returns "3M"
abbrev_format(42.7, None)  # Returns "42.7"

display_sliders(name: str, sliders: mo.ui.dictionary | mo.ui.slider, invars: dict[str, dict], dist: str | Callable | None = None, descriptions: dict = {}) -> mo.Html

Display parameter sliders with optional distribution plot.

Parameters:

• name: Name of the parameter group to display
• sliders: Either a single slider or dictionary of sliders for distribution params
• invars: Dictionary to store input variable configurations
• dist: Distribution to use (string name or callable), required for multi-sliders
• descriptions: Optional dict mapping parameter names to descriptions

Returns:

• Marimo component displaying the sliders and optional distribution plot

For a single slider, displays it as a constant parameter. For multiple sliders, displays them with a plot of the resulting distribution. Distribution must be specified for multiple sliders, either as a string name matching a scipy distribution or as a callable distribution object.

Raises:

• DistributionConfigurationError: If dist is None for multiple sliders

generate_ranges(distribution: str, ranged_distkwargs: dict) -> dict

Generate parameter ranges for a probability distribution.

Takes a distribution name and dictionary of parameter ranges, validates the ranges against allowed bounds for that distribution, and returns a merged dictionary with complete range specifications.

Parameters:

• distribution: Name of the probability distribution (e.g. "normal", "beta")
• ranged_distkwargs: Dictionary mapping parameter names to their range specifications. Each range spec should have "lower" and "upper" bounds.

Returns:

• dict: Complete parameter range specifications with distribution defaults merged with provided ranges

Raises:

• MissingParameterError: If a required parameter range is not provided
• ParameterBoundError: If a provided range exceeds the allowed bounds for a param

params_sliders(ranged_distkwargs: dict) -> mo.ui.dictionary

Create a dictionary of sliders for parameter ranges.

Takes a dictionary of param ranges and creates interactive sliders for each param. The sliders will be bounded by the lower/upper values specified in the ranges dict. The step size and initial value can optionally be specified per param.

Parameters:

• ranged_distkwargs: Dictionary mapping parameter names to their range specifications. Each range spec should have "lower" and "upper" bounds, and optionally "step" and "value" keys.

Returns:

• mo.ui.dictionary: A dictionary of marimo slider UI elements, one per parameter. Each slider will be configured according to the parameter's range spec.

Example:

ranges = {
    "mean": {"lower": 0, "upper": 100, "step": 1, "value": 50},
    "std": {"lower": 0, "upper": 10}
}
sliders = params_sliders(ranges)

Supported Distributions

Any scipy distribution can be used.

Included helpers following scipy distributions:

• uniform: Uniform distribution
• triang: Triangular distribution
• beta: Beta distribution
• norm: Normal distribution

Dependencies

• marimo
• matplotlib
• scipy