Easy memoization for Julia.
using Memoize
@memoize function x(a)
println("Running")
2a
end
julia> x(1)
Running
2
julia> memoize_cache(x)
IdDict{Any,Any} with 1 entry:
(1,) => 2
julia> x(1)
2
julia> empty!(memoize_cache(x))
IdDict{Any,Any}()
julia> x(1)
Running
2
julia> x(1)
2
By default, Memoize.jl uses an IdDict
as a cache, but it's also possible to specify the type of the cache. If you want to cache vectors based on the values they contain, you probably want this:
using Memoize
@memoize Dict function x(a)
println("Running")
a
end
You can also specify the full function call for constructing the dictionary. For example, to use LRUCache.jl:
using Memoize
using LRUCache
@memoize LRU{Tuple{Any,Any},Any}(maxsize=2) function x(a, b)
println("Running")
a + b
end
julia> x(1,2)
Running
3
julia> x(1,2)
3
julia> x(2,2)
Running
4
julia> x(2,3)
Running
5
julia> x(1,2)
Running
3
julia> x(2,3)
5
Note that the @memoize
macro treats the type argument differently depending on its syntactical form: in the expression
@memoize CacheType function x(a, b)
# ...
end
the expression CacheType
must be either a non-function-call that evaluates to a type, or a function call that evaluates to an instance of the desired cache type. Either way, the methods Base.get!
and Base.empty!
must be defined for the supplied cache type.