This software is released under MIT license. 本软件在MIT许可下发行

This module provides uncertainty class that helps address error calculations in college-level courses like Basic Physics Lab and Physical Chemistry Lab.

该模块提供uncertainty（不确定度）类，帮助解决各高校《基础物理实验》（基物实验）和《物理化学实验》（物化实验）等课程中的误差计算。

Functions功能

1. Provide a class called uncertainty that has three attributes: value, error, name (optional)
2. Error is calculated on the fly, and $\LaTeX$ expression that lead to the error is easily generated on request.
3. Each variable is created independently so that the correlation induced by repeated presence of the same variable into account.
4. This class is compatible as a dtype with vector operation classes like numpy.ndarray, pandas.DataFrame, so that large amount of data can be processed.
5. Supports all scientific operations: power, exponential, logarithm, trigonometric function, etc.
6. Operations that introduce additional error between multiple repeated measurements are implemented: mean and least-square linear regression.

1. 提供一个名为 uncertainty（不确定度） 的类，该类具有三个属性：value（值）、error（误差）、name（名称，可选）
2. 动态计算误差，并且可以根据要求轻松生成推导误差的 $\LaTeX$ 表达式。
3. 每个变量都是独立创建的，以考虑由相同变量重复存在引起的相关性。
4. 此类可作为dtype与numpy.ndarray、pandas.DataFrame等向量操作类兼容，以处理大量数据。
5. 支持所有科学运算：幂、指数、对数、三角……
6. 实现了在多次重复测量之间引入额外误差的操作：均值和最小二乘线性回归。

Basic Usage用法简介

1. Download uncertainty.py and place it under your data folder.
2. Open Python in that folder and input from uncertainty import uncertainty
3. Use uncertainty(value, error, name) to create an uncertainty object "name = value ± error". Use float as constant.
4. Call .latex() method on object to generate $\LaTeX$ expression.
5. uncertainty.mean, uncertainty.lsmr are two class functions that allows you to consider the extra error introduced by multiple measurements.
6. Try using uncertainty as dtype in NumPy and pandas in order to process large amount of data.
7. Refer to doc by Python help whenever necessary. If there is bug, please report to me: zsc_chris@outlook.com. I'll try to solve it with you.

1. 下载uncertainty.py并将其放在您的数据文件夹下。
2. 在该文件夹中打开Python并输入from uncertainty import uncertainty
3. 使用uncertainty(value, error, name)创建一个uncertainty对象name = value ± error。使用浮点数表示常数。
4. 在对象上调用.latex()方法生成 $\LaTeX$ 表达式。
5. uncertainty.mean、uncertainty.lsmr是两个类函数，可让您考虑多次测量引入的额外误差。
6. 尝试在NumPy和pandas中使用uncertainty作为dtype，以处理大量数据。
7. 有任何需求，请用Python help功能查看文档。如果有bug，请联系我：zsc_chris@outlook.com。我会尝试和您一起解决。

Requirements依赖

torch*, sympy

*just CPU version is OK只要CPU版本即可

Acknowledgement致谢

Thank the teaching assistant Junhan Chang of "Machine Learning and its Applications in Chemistry" course for giving the inspiration of using PyTorch as autograd tool. Thank Zhaoyang Li, a top senior student in last grade, whose single-formula $\LaTeX$ generator UncertaintyCalculator provides me the motivation to develop a more flexible $\LaTeX$ generator protocol.

感谢《机器学习及其在化学中的应用》课程的昌珺涵助教给了我使用PyTorch作为自动求导工具的启发。感谢比我高一级的学霸李昭阳学长，他的单公式 $\LaTeX$ 生成器UncertaintyCalculator给了我开发更灵活的 $\LaTeX$ 生成方式的动力。