Bijections
This package provides a Bijection
data type for Julia.
A Dict
in Julia is not onetoone. Two different keys might have the
same value. This data structure behaves just like a Dict
except it
blocks assigning the same value to two different keys.
Getting Started
After using Bijections
we create a new Bijection
in one of the
following ways:

b = Bijection()
: This gives a newBijection
in which the keys and values are ofAny
type. 
b = Bijection{S,T}()
: This gives a newBijection
in which the keys are of typeS
and the values are of typeT
. 
b = Bijection(x,y)
: This gives a newBijection
in which the keys are typetypeof(x)
, the values are typetypeof(y)
and the keyvalue pair(x,y)
is inserted into theBijection
. 
b = Bijection(dict::AbstractDict{S, T})
: This gives a newBijection
in which the keys are typeS
, the values are typeT
and all keyvalue pairs indict
are inserted into theBijection
.
Adding and deleting pairs
Once a Bijection
, b
, is created, we add a new keyvalue pair in
the same manner as with a Dict
:
julia> b[1] = "hello"
"hello"
julia> b[2] = "bye"
"bye"
Notice, however, that if we add a new key with a value that already
exists in the Bijection
an error ensues:
julia> b[3] = "hello"
ERROR: One of x or y already in this Bijection
Likewise, if a key already has a value it cannot be changed by giving it a new value:
julia> b[1] = "ciao"
ERROR: One of x or y already in this Bijection
If we wish to change the value associated with a given key, the pair
must first be deleted using delete!
:
julia> delete!(b,1)
Bijection{Any,Any} (with 1 pairs)
julia> b[1] = "ciao"
"ciao"
Bijection
Using a To access a value associated with a given key, we use the same syntax
as for a Dict
:
julia> b[1]
"ciao"
julia> b[2]
"bye"
If the key is not in the Bijection
an error is raised:
julia> b[3]
ERROR: KeyError: 3 not found
Since the values in a Bijection
must be distinct, we can give a
value as an input and retrieve its associate key. The function
inverse(b,y)
finds the value x
such that b[x]==y
. However, we
provide the handy short cut b(y)
:
julia> b("bye")
2
julia> b("ciao")
1
Naturally, if the requested value is not in the Bijection
an error
is raised:
julia> b("hello")
ERROR: KeyError: hello not found
Bijection
Creating an inverse There are two functions that take a Bijection
and return a new
Bijection
that is the functional inverse of the original:
inv
and active_inv
.
inv
Independent inverse: Given a Bijection
b
, calling inv(b)
creates a new Bijection
that is the inverse of b
. The new Bijection
is completely independent
of the original, b
. Changes to one do not affect the other:
julia> b = Bijection{Int,String}()
Bijection{Int64,String} (with 0 pairs)
julia> b[1] = "alpha"
"alpha"
julia> b[2] = "beta"
"beta"
julia> bb = inv(b)
Bijection{String,Int64} (with 2 pairs)
julia> bb["alpha"]
1
julia> bb["alpha"]
1
julia> b[3] = "gamma"
"gamma"
julia> bb["gamma"]
ERROR: KeyError: key "gamma" not found
active_inv
Active inverse: The active_inv
function also creates an inverse Bijection
, but in this
case the original and the inverse are actively tied together.
That is, modification of one immediately affects the other.
The two Bijection
s remain inverses no matter how either is modified.
julia> b = Bijection{Int,String}()
Bijection{Int64,String} (with 0 pairs)
julia> b[1] = "alpha"
"alpha"
julia> b[2] = "beta"
"beta"
julia> bb = active_inv(b)
Bijection{String,Int64} (with 2 pairs)
julia> b[3] = "gamma"
"gamma"
julia> bb["gamma"]
3
Iteration
Bijection
s can be used in a for
statement just like Julia
dictionaries:
julia> for (x,y) in b; println("$x > $y"); end
2 > beta
3 > gamma
1 > alpha
Inspection
Thinking of a Bijection
as a mapping between finite sets, we
provide the functions domain
and image
. These return,
respectively, the set of keys and the set of values of the
Bijection
.
julia> domain(b)
Set(Any[2,1])
julia> image(b)
Set(Any["bye","ciao"])
The collect
function returns the Bijection
as an array of
keyvalue pairs:
julia> collect(b)
2element Array{Tuple{Any,Any},1}:
(2,"bye")
(1,"ciao")
The length
function returns the number of keyvalue pairs:
julia> length(b)
2
The isempty
function returns true
exactly when the Bijection
contains no pairs:
julia> isempty(b)
false