UncValue library released

Today, together with Àlex Giménez Romero we have released the UncValue library for Python and Julia. This library allows to initialize a number with an uncertainty and propagate that uncertainty under all the operations performed.

For instance, imagine you measured a rectangular table whose sides measure 1.5m and 80cm with ruler with precision up to 1mm. Then, you can initialize the Value as

• Python:

  from uncvalue import Value
  L1 = Value(1.5, 1e-3)
  L2 = Value(0.8, 1e-3)
  

• Julia:

  using UncValue
  L1 = Value(1.5, 1e-3)
  L2 = Value(0.8, 1e-3)
  

You now want to calculate the area of the table, so you multiply both lengths

• Python:

  A = L1 * L2
  print(A)
  

• Julia:

  A = L1 * L2
  println(A)
  

  and obtain as outcome (1200.0 ± 1.7)·10^-3.

Of course, it is possible to perform more complex operations like

• Python (numpy required, functions from python math will only compute the value)

  import numpy as np
  print(L1 ** L2)  # power -> (13831.6 ± 9.3)·10^-4
  print(np.sin(L1))  # sinus -> (99749.5 ± 7.1)·10^-5
  print(np.exp(L2))  # exponential -> (2225.5 ± 2.2)·10^-3
  

• Julia

  println(L1^L2)  # power -> (13831.6 ± 9.3)·10^-4
  println(sin(L1))  # sinus -> (99749.5 ± 7.1)·10^-5
  println(exp(L2))  # exponential -> (2225.5 ± 2.2)·10^-3
  

For more information visit the original repositories:

• Python: source code, docs
• Julia: source code