#!/usr/bin/env python3
# -*- coding: utf-8 -*-
# Time-stamp: "2025-06-04 11:05:00 (ywatanabe)"
# File: ./src/scitex/str/_factor_out_digits.py
"""
Functionality:
Factor out common powers of 10 from numerical data for cleaner axis labels
Input:
Numerical data (list, array, or individual numbers)
Output:
Factored values and the common factor for display
Prerequisites:
numpy
"""
import numpy as np
from typing import Union, List, Tuple, Optional
[docs]
def factor_out_digits(
values: Union[List, np.ndarray, float, int],
precision: int = 2,
min_factor_power: int = 3,
return_latex: bool = True,
return_unicode: bool = False
) -> Tuple[Union[List, np.ndarray, float], str]:
"""
Factor out common powers of 10 from numerical values for cleaner scientific notation.
Parameters
----------
values : Union[List, np.ndarray, float, int]
Numerical values to factor
precision : int, optional
Number of decimal places in factored values, by default 2
min_factor_power : int, optional
Minimum power of 10 to factor out, by default 3
return_latex : bool, optional
Return factor string in LaTeX format, by default True
return_unicode : bool, optional
Return factor string with Unicode superscripts, by default False
Returns
-------
Tuple[Union[List, np.ndarray, float], str]
Tuple of (factored_values, factor_string)
Examples
--------
>>> factor_out_digits([1000, 2000, 3000])
([1.0, 2.0, 3.0], '$\\times 10^{3}$')
>>> factor_out_digits([0.001, 0.002, 0.003])
([1.0, 2.0, 3.0], '$\\times 10^{-3}$')
>>> factor_out_digits([1.5e6, 2.3e6, 4.1e6])
([1.5, 2.3, 4.1], '$\\times 10^{6}$')
"""
# Convert to numpy array for easier processing
if np.isscalar(values):
values_array = np.array([values])
is_scalar = True
else:
values_array = np.array(values)
is_scalar = False
# Remove zeros and handle special cases
non_zero_values = values_array[values_array != 0]
if len(non_zero_values) == 0:
return values, ""
# Find the common order of magnitude
log_values = np.log10(np.abs(non_zero_values))
common_power = int(np.floor(np.mean(log_values)))
# Only factor out if the power is significant enough
if abs(common_power) < min_factor_power:
return values, ""
# Calculate factored values
factor = 10 ** common_power
factored_values = values_array / factor
# Round to specified precision
factored_values = np.round(factored_values, precision)
# Generate factor string
factor_string = _format_factor_string(common_power, return_latex, return_unicode)
# Return in original format
if is_scalar:
return float(factored_values[0]), factor_string
else:
if isinstance(values, list):
return factored_values.tolist(), factor_string
else:
return factored_values, factor_string
[docs]
def auto_factor_axis(
ax,
axis: str = 'both',
precision: int = 2,
min_factor_power: int = 3,
return_latex: bool = True,
return_unicode: bool = False,
label_offset: Tuple[float, float] = (0.02, 0.98)
) -> None:
"""
Automatically factor out common powers of 10 from axis tick labels.
Parameters
----------
ax : matplotlib.axes.Axes
The axes object to modify
axis : str, optional
Which axis to factor ('x', 'y', or 'both'), by default 'both'
precision : int, optional
Number of decimal places in factored values, by default 2
min_factor_power : int, optional
Minimum power of 10 to factor out, by default 3
return_latex : bool, optional
Use LaTeX format for factor string, by default True
return_unicode : bool, optional
Use Unicode superscripts for factor string, by default False
label_offset : Tuple[float, float], optional
Position for factor label as (x, y) in axes coordinates, by default (0.02, 0.98)
Examples
--------
>>> import matplotlib.pyplot as plt
>>> fig, ax = plt.subplots()
>>> ax.plot([1000, 2000, 3000], [0.001, 0.002, 0.003])
>>> auto_factor_axis(ax, axis='both')
"""
if axis in ['x', 'both']:
_factor_single_axis(ax, 'x', precision, min_factor_power,
return_latex, return_unicode, label_offset)
if axis in ['y', 'both']:
_factor_single_axis(ax, 'y', precision, min_factor_power,
return_latex, return_unicode, label_offset)
def _factor_single_axis(
ax,
axis_name: str,
precision: int,
min_factor_power: int,
return_latex: bool,
return_unicode: bool,
label_offset: Tuple[float, float]
) -> None:
"""Factor out digits from a single axis."""
# Get current tick values
if axis_name == 'x':
tick_values = ax.get_xticks()
set_ticks = ax.set_xticks
set_ticklabels = ax.set_xticklabels
label_x, label_y = label_offset[0], 0.02
else: # y-axis
tick_values = ax.get_yticks()
set_ticks = ax.set_yticks
set_ticklabels = ax.set_yticklabels
label_x, label_y = 0.02, label_offset[1]
# Factor out common digits
factored_ticks, factor_string = factor_out_digits(
tick_values, precision, min_factor_power, return_latex, return_unicode
)
# Apply factored labels if factor was found
if factor_string:
set_ticks(tick_values)
set_ticklabels([f"{val:.{precision}f}" for val in factored_ticks])
# Add factor label
ax.text(label_x, label_y, factor_string,
transform=ax.transAxes,
fontsize='small',
bbox=dict(boxstyle="round,pad=0.3", facecolor='white', alpha=0.8))
def _format_factor_string(
power: int,
latex: bool = True,
unicode_sup: bool = False
) -> str:
"""
Format the factor string for display.
Parameters
----------
power : int
The power of 10
latex : bool, optional
Use LaTeX formatting, by default True
unicode_sup : bool, optional
Use Unicode superscripts, by default False
Returns
-------
str
Formatted factor string
"""
if latex:
return f"$\\times 10^{{{power}}}$"
elif unicode_sup:
# Unicode superscript mapping
superscripts = {
'0': '⁰', '1': '¹', '2': '²', '3': '³', '4': '⁴',
'5': '⁵', '6': '⁶', '7': '⁷', '8': '⁸', '9': '⁹',
'-': '⁻', '+': '⁺'
}
power_str = str(power)
unicode_power = ''.join(superscripts.get(char, char) for char in power_str)
return f"×10{unicode_power}"
else:
return f"×10^{power}"
# EOF