martyo256/py-formatted-value

Simple Python package to format values and their uncertainties up to a given amount of significant figures in the uncertainty.

py-formatted-value

---

Format experimental values and their uncertainties using appropriate significant figures and matching decimal places.

---

Installation

To install fvalue, run:

pip install fvalue

Examples

The following examples showcase the core features of the library.

from fvalue import FormattedValue
from decimal import Decimal

FormattedValue(
    value=10_973_731.768_160,
    error=0.000_021,
    error_significant_figures=2,
    rounding=FormattedValue.RoundingOption.ROUND_HALF_EVEN,
).formatted(
    template=FormattedValue.SIUNITX_TEMPLATE,
    units=r"\per\meter",
)
# >>> "\SI{10973731.768160 \pm 0.000021 e0}{\per\meter}"

FormattedValue(
    value=0.000_002_671,
    error=0.000_000_452,
    error_significant_figures=1,
).formatted(
    template=FormattedValue.SIUNITX_NUM_TEMPLATE,
)
# >>> "\num{2.7 \pm 0.5 e-6}"

FormattedValue(
    0.000_002_671,
    0.000_000_452,
    error_significant_figures=2,
).formatted(
    FormattedValue.SIUNITX_NUM_TEMPLATE,
    multiplier=10 ** 3,
)
# >>> "0.00267 ± 0.00045"

FormattedValue(
    Decimal("1.602_176_634E-19"),
).formatted(
    FormattedValue.SIUNITX_VALUE_TEMPLATE,
    units=r"\coulomb",
)
# >>> "\SI{1.602176634 e-19}{\coulomb}"

FormattedValue(
    Decimal("1.416_784E32"),
    Decimal("0.000_016E32"),
    error_significant_figures=2,
).formatted(
    FormattedValue.NATURAL_TEMPLATE,
    units=r"K",
)
# >>> "(1416784 ± 16) x 10^26 K"

FormattedValue(
    656,
    10,
).formatted(
    FormattedValue.NATURAL_TEMPLATE,
    units=r"nm",
)
# >>> "(66 ± 1) x 10 nm"

Usage

Instances of FormattedValue are used to produce string representations of experimental quantities with their uncertainties. The number of significant figures in the error on a value is set on instantiation. The decimal places of the value always match that of the error in formatted strings. Values of small magnitude are automatically formatted with an appropriate decimal exponent in the scientific notation.

Rounding

The rounding policies used in the formatting are those of the decimal library, namely: ROUND_HALF_EVEN, ROUND_05UP, ROUND_CEILING, ROUND_DOWN, ROUND_FLOOR, ROUND_HALF_DOWN, ROUND_HALF_UP and ROUND_UP. All of these are conveniently available as an enumeration at FormattedValue.RoundingOption. By default, ROUND_HALF_EVEN is used to mitigate some biases.

Templates

Templates allow for the placement in strings of the rounded value, error, decimal exponent, and units optionally. Using the formatted method of an instance of FormattedValue, strings and callable templates can be used.

If a given template is a string, then:

• {0} corresponds to the rounded value;
• {1} corresponds to the rounded error;
• {2} corresponds to the decimal exponent in scientific notation;
• {3} corresponds to the units.

For instance, "({0} ± {1}) x 10^{2} {3}" would generate "(10 ± 1) x 10^0 m" for a formatted value with value 10, error 1, and units "m".

If a given template is a function, then it should have type signature Callable[[str, str, str, str], str], with arguments (value: str, error: str, exponent: str, units: str) -> str. One such template is accessible at FormattedValue.NATURAL_TEMPLATE.

Various templates are available to speed up the generation of formatted strings. The default template is for SIUNITX, which is accessible at FormattedValue.SIUNITX_TEMPLATE and produces strings as "\SI{{0} \pm {1} e{2}}{{3}}". Other templates include: Formatted.SIUNITX_VALUE_TEMPLATE, Formatted.SIUNITX_ERROR_TEMPLATE, Formatted.SIUNITX_NUM_TEMPLATE, Formatted.SIUNITX_NUM_VALUE_TEMPLATE, Formatted.SIUNITX_NUM_ERROR_TEMPLATE.

Multiplier

The formatted method of instances of FormattedValue allows for a multiplier to be applied to both the value and error of the instance. This allows for an experimental value to be used using the International System of Units for data processing, and prefixed units to be used for the presentation of results in reports.

About

Authors

• Marc-Antoine Ouimet - MartyO256

License

This project is licensed under the MIT License. See the LICENSE.md file for details.