dabacon / blqs-1

Python building blocks for quantum computing and domain specific languages.

blqs: Building Blocks for Domain Specific Languages

blqs: blqs_cirq:

Blqs (pronounced "blocks") is a framework for building (internal) domain specific languages in Python. It was inspired by TensorFlow's autograph library, and motivated by attempting to improve quantum computing programming frameworks like Cirq and Qiskit. If you want to write an domain specific language that uses native python code for "if" statements or "for" statements or "while" statements, blqs is for you!

Packages

This repo contains two packages:

• blqs The base blqs framework (installation instructions, introductory documentation).
• blqs_cirq An application of blqs to Cirq (installation instructions, introductory documentation).

Motivating example

Here is a motivating example. In many traditional quantum programming frameworks one writes a quantum program via appending to a container object, (here is an example from Cirq):

import cirq

circuit = cirq.Circuit()
q0, q1 = cirq.LineQubit.range(2)
circuit.append(cirq.H(q0))
circuit.append(cirq.CX(q0, q1))

In blqs instead one can define functions that return programs:

import blqs

H = blqs.Op('H')
CX = blqs.Op('CX')

@blqs.build
def hello_blqs():
    H(0)
    CX(0, 1)

program = hello_blqs()
print(program)
> prints
> H 0
> H 1

Here program is blqs.Program container of the listed blqs.Statements. The function annotation turns the hello_blqs function into a builder that, when called, returns the built program.

More interestingly, blqs programs can also take native python functionality, like if statements, and capture them in blqs objects:

M = blqs.Op('M')

@blqs.build
def hello_if_blqs():
    a = blqs.Register('a')
    H(0)
    M(0, 'a')
    if a:
        CX(0, 1)
    else:
        CX(1, 0)

program = hello_if_blqs()
print(program)
> prints
> H 0
> M 0, a
> if a:
>     CX 0, 1
> else:
>     CX 1, 0

# This is three statements, the last having captured the if.
for s in program:
    print(type(s))
> prints
> <class 'blqs.instruction.Instruction'>
> <class 'blqs.instruction.Instruction'>
> <class 'blqs.conditional.If'>

Further we can mix and match native and captured Python. For example, here a is just a normal python boolean variable, and we can use it in to build one of two different cases:

@blqs.build
def hello_native(a):
    if a:
        H(0)
    else:
        H(1)
print(hello_native(True))
> prints
> H 0
print(hello_native(False))
> prints
> H 1

Blqs is meant to be all about building programs and the intermediate representation of that program. In many ways it is meant to be a framework to help you build other frameworks. As such, every attempt will be made to keep blqs simple so that other frameworks can be built on top of it.

Contributing

We welcome contributions! In order to contribute we require that a Contributor License Agreement (CLA) be signed by the contributor or any organization (business, school) that retains rights to your contributed code. To get setup with a development environment and for info on the CLA see here.

License

Both blqs and blqs_cirq are licensed under Apache 2.0.