NamedTrajectories.jl

Efficient Handling of Trajectories with User Defined Named Components
Author kestrelquantum
Popularity
6 Stars
Updated Last
2 Months Ago
Started In
January 2023

NamedTrajectories.jl

Documentation Build Status Support
Stable Dev Build Status Coverage Unitary Fund

NamedTrajectories.jl is a package for working with trajectories of named variables. It is designed to be used with QuantumCollocation.jl and Piccolo.jl.

Please see the dev documentation for more information.

Installation

NamedTrajectories.jl is now registered! Install in the REPL by entering pkg mode with ] and then running

pkg> add NamedTrajectories

or to install the latest master branch run

pkg> add NamedTrajectories#main

Basic Usage

Users can define NamedTrajectory types which have lots of useful functionality. For example, you can access the data by name or index. In the case of an index, a KnotPoint is returned which contains the data for that timestep.

using NamedTrajectories

# define number of timesteps and timestep
T = 10
dt = 0.1

# build named tuple of components and data matrices
components = (
    x = rand(3, T),
    u = rand(2, T),
)

# build trajectory
traj = NamedTrajectory(components; timestep=dt, controls=:u)

# access data by name
traj.x # returns 3x10 matrix of x data
traj.u # returns 2x10 matrix of u data

z1 = traj[1] # returns KnotPoint with x and u data

z1.x # returns 3 element vector of x data at timestep 1
z1.u # returns 2 element vector of u data at timestep 1

traj.data # returns data as 5x10 matrix
traj.names # returns names as tuple (:x, :u)

# simple build in plotting
plot(traj)

This script generates the following plot as demonstrated in demo.ipynb

TODO

  • Add contributing guidelines for
    • documentation
    • new features
    • testing
    • examples
  • Flesh out manual in documentation, see docs/literate/man or docs/literate/manual.jl
  • Convert literate files to notebooks for examples
  • Internal optimizations
    • Handling of data as views
    • KnotPoint as view
    • More efficient copying and constructing
  • Docstrings for all functions
  • functionality
    • random NTs