Containers

Containers

Module defining containers for conveniently working with filament data.

VectorOfVectors{T, V <: AbstractVector{T}} <: AbstractVector{V <: AbstractVector}
VectorOfVectors(data::AbstractVector{<:AbstractVector})

Contains a list of vectors.

It behaves as much as possible as a basic vector of vectors. For instance:

• its length is the number of contained vectors;

• its element type (eltype) is the type of a contained vector;

• it only allows linear indexing (e.g. u[i] to get a single vector). To get an individual element (of type T), one should do u[i][j].

Note that the individual vectors u[i] can have different lengths from each other.

There are also some differences:

• it overloads broadcasting mechanisms, so that doing @. u = 2 * u + v (where both variables are VectorOfVectors) works as expected and is efficient;

• copying a VectorOfVectors recursively copies its contained vectors, instead of just copying array references ("pointers").

Examples

julia> data = [rand(n) for n ∈ 1:4]
4-element Vector{Vector{Float64}}:
 [0.6293451231426089]
 [0.4503389405961936, 0.47740714343281776]
 [0.7031298490032014, 0.6733461456394962, 0.16589443479313404]
 [0.6134782250008441, 0.6683403279577278, 0.4570310908017041, 0.2993652953937611]

julia> us = VectorOfVectors(data)
4-element VectorOfVectors{Float64, Vector{Float64}}:
 [0.6293451231426089]
 [0.4503389405961936, 0.47740714343281776]
 [0.7031298490032014, 0.6733461456394962, 0.16589443479313404]
 [0.6134782250008441, 0.6683403279577278, 0.4570310908017041, 0.2993652953937611]

julia> us[2]
2-element Vector{Float64}:
 0.4503389405961936
 0.47740714343281776

julia> vs = @. us + 2 * us  # broadcasting
4-element VectorOfVectors{Float64, Vector{Float64}}:
 [1.888035369427827]
 [1.3510168217885807, 1.4322214302984533]
 [2.109389547009604, 2.0200384369184885, 0.4976833043794021]
 [1.8404346750025324, 2.0050209838731834, 1.3710932724051124, 0.8980958861812833]