AutoVectors.jl

Julia vectors whose first and last indices are any integers

Most languages make a choice: arrays start at 0 or 1. For some applications, one would like arrays that start and end anywhere, e.g. the array runs from -3 to 10. AutoVectors have this property. They also auto-resize based on writing outside their range. The range is always contigous, and implemented through ordinary vectors, so AutoVectors are fast, say, compared to a Dict{Int64,Float64} which might implement similar featues. Reading outside the range gives 0.0 (or whatever the default element is). Often these properties mean that you don't have to worry about what the range is; it just works.

For example

julia> v = AutoVector(0.0)        # default element, determining type

AutoVector{Float64}(1, 0, 0, 0.0, Float64[])

julia> v[-3] = pi

π = 3.1415926535897...

julia> v[10] = exp(1.0)

2.718281828459045

julia> doprint(v)

-3 3.141592653589793

-2 0.0

-1 0.0

0 0.0

1 0.0

2 0.0

3 0.0

4 0.0

5 0.0

6 0.0

7 0.0

8 0.0

9 0.0

10 2.718281828459045

julia> avrange(v)

-3:10

julia> v[15]      # No resizing, give default

0.0

julia> avrange(v)

-3:10

You can add and subtract AutoVectors (e.g. v+w, v-w), add or subtract a constant to all elements (v+a,v-a),
and multiply by a constant, a*v or v*a, and v/a. Broadcasting (v .* w) mostly doesn't work, since the lengths are generally different, but there is a function pointmult(v,w) which does the same thing. A number of other useful functions are also implemented, such as dot products and convolutions.

For full documentation, see documentation