README.md

Quantikz.jl

A minimal package for drawing quantum circuits using the quantikz and tikz TeX libraries.

To install it use:

] add Quantikz

The library can generate tex, pdf, and png files, as well as live previews. It does not require anything to be installed on your system (the tex and image manipulation dependencies are handled by Julia, with Tectonic.jl and FileIO.jl). If you want to generate another image type, simply use FileIO.

circuit = [CNOT(1,2),Measurement(2)]
displaycircuit(circuit) # you can set a scale parameter

To save a png/pdf file:

savecircuit(circuit, "file.png") # you can set a scale parameter
savecircuit(circuit, "file.pdf")

You can view the corresponding TeX string or save it to file with:

circuit2string(circuit)
savetex(circuit, "file.tex")

Example

circuit = [
    CNOT(1,2), CPHASE(2,3), SWAP(3,4),
    H(5), P(6), Id(7),
    U("\\frac{\\phi}{4}",8),
    Measurement("X",1), Measurement([2,3],2), ClassicalDecision("U",[3,5],2),
    Measurement("M",[5,6],1),
    MultiControlU("G",[2,8],[7,3],[4,5,6])]

In standard mode:

And expanded for easier identification:

Built-in quantum circuit operations

CNOT, CPHASE, SWAP, H, P, Id, a generic single qubit gate U, a generic measurement Measurement, a classically conditioned gate ClassicalDecision, and a parity check measurement ParityMeasurement.

The general purpose MultiControl(controls, ocontrols, targets, targetXs) can be used to create an arbitrary combination of muticontrol multiqubit gates. Each argument lists the indices of qubits that should get a certain symbol: controls is filled circles, ocontrols is empty circles, targets is the NOT symbol, and targetXs is the X symbols.

For named controlled gates use MultiControlU(str, controls, ocontrols, targets).

For noise events, you can use Noise(targets) or NoiseAll().

For examples of these operations, see the complete list.

Custom objects

For your CustomQuantumOperation simply define a QuantikzOp(op::CustomQuantumOperation) that converts your object to one of the built-in objects.

If you need more freedom for your custom quantum operation, simply define:

• update_table!(table,step,op::CustomQuantumOperation) that directly modifies the quantikz table
• affectedqubits(op::CustomQuantumOperation) that gives the indices of qubits involved in the operation.
• (optional) affectedbits(op) that gives the indices of classical bits in use (empty by default)
• (optional) neededancillaries(op) that gives the number of temporary ancillary qubits to reserve (0 by default)
• (optional) nsteps(op) that gives the number of steps involved in the gate (1 by default)
• (optional) deleteoutputs(op) that gives which qubits to be deleted, e.g., their lines removed (empty by default)

Instead of returning an array of indices affectedqubits can also return the lazy slice ibegin:iend (from EndpointRanges.jl) which tells the layout engine that all qubits are used in this stage of the circuit.

Internally, this library converts the array of circuit operations to a 2D array of quantikz macros which is then converted to a single TeX string, which is then compiled with a call to Tectonic.jl.

Under the hood

If you need some of the more advanced features of the quantikz TeX macros that are not implemented here yet, you can edit the string directly, or more conveniently, you can generate the 2D array of macros that makes up the string:

julia> circuit2table([CNOT(1,2),CNOT(2,3)])
3×4 Array{String,2}:
 "\\qw"  "\\ctrl{1}"  "\\qw"       "\\qw"
 "\\qw"  "\\targ{}"   "\\ctrl{1}"  "\\qw"
 "\\qw"  "\\qw"       "\\targ{}"   "\\qw"

table2string (and string2image) can be used for the final conversion.

LaTeX-free alternatives

An alternative is the YaoPlots.jl which can draw directly from Julia (does not shell out to ImageMagick and Ghostscript). YaoPlots.jl however does not support LaTeX.