qiskitflow 0.0.4a0

QiskitFlow. Reproducible quantum experiments.

QiskitFlow. Reproducible quantum experiments.

Platform for tracking, sharing and running quantum experiments in a clean and understandable for developers, researchers and students manner.

Alpha release is in the works. Files from hackathon project are located in hackathon folder

General overview

Hot to run

• Overview / Flow
• Installation
• Code annotation
• CLI
  • Experiments list
  • Experiment information
• Examples
• UI Mocks

Flow

Flow of actions while using QiskitFlow is following:

• Install QiskitFlow if not installed yet
• Annotate your code with Experiment abstraction QiskitFlow library provides
• Run your code as usual: QiskitFlow will write metadata of your experiment execution in local folder
• You can review experiments using CLI interface

We are tracking metrics, parameters, artifacts and measurements of experiments. Artifacts, circuits, sourcecode and other useful things are on their way.

Note: qiskitflow creates _experiments folder in place of execution of code, where all serialized information is stored in json format, so it's easy to track it even in git

Installation

pip install qiskitflow

Code annotation

Library for quantum programs annotation

Sample example of annotation:

from qiskitflow import Experiment

with Experiment("awesome_experiment") as experiment:

    # your quantum program here!

    experiment.write_metric("test metric", 0.1)
    experiment.write_metric("test metric 2", 2)

    experiment.write_parameter("test parameter", "test parameter value")
    experiment.write_parameter("test parameter 2", "test paraeter value 2")

    experiment.write_counts("measurement", {"00": 1024, "11": 0})

Full example with quantum teleportation

import numpy as np
import time
from qiskit import QuantumCircuit, QuantumRegister, ClassicalRegister, execute, BasicAer, IBMQ
from qiskit.visualization import plot_histogram, plot_bloch_multivector
from qiskit.extensions import Initialize
from qiskit_textbook.tools import random_state, array_to_latex

from qiskitflow import Experiment

with Experiment("quantum teleportation") as experiment:
    start_time = time.time()

    # conduct experiment as usual
    psi = random_state(1)
    init_gate = Initialize(psi)
    init_gate.label = "init"
    inverse_init_gate = init_gate.gates_to_uncompute()

    qr = QuantumRegister(3, name="q")
    crz = ClassicalRegister(1, name="crz")
    crx = ClassicalRegister(1, name="crx")
    qc = QuantumCircuit(qr, crz, crx)
    qc.append(init_gate, [0])
    qc.barrier()
    create_bell_pair(qc, 1, 2)
    qc.barrier()
    alice_gates(qc, 0, 1)
    measure_and_send(qc, 0, 1)
    bob_gates(qc, 2, crz, crx)
    qc.append(inverse_init_gate, [2])
    cr_result = ClassicalRegister(1)
    qc.add_register(cr_result)
    qc.measure(2,2)
    backend_name = "qasm_simulator"
    backend = BasicAer.get_backend(backend_name)
    counts = execute(qc, backend, shots=1024).result().get_counts()

    end_time = time.time()

    runtime = end_time - start_time

    # qiskitflow =========

    # log parameters used
    experiment.write_parameter("backend name", backend_name)
    # log metrics of experiment
    experiment.write_metric("runtime", runtime)
    # log counts of experiment
    experiment.write_counts("experiment counts", counts)

---

CLI

List of experiments

qiskitflow experiments list

experiments list screenshot

Experiment information (runs)

qiskitflow experiments info <NAME_OF_EXPERIMENT> --metrics="<METRIC_NAME>,<OTHER_METRIC_NAME>" --parameters="<PARAM>,<OTHER PARAM>"

experiment information screenshots

Experiment information

Experiment with specified optional flags

---

Examples

Jupyter notebook with quantum teleportation example

Mocks of upcoming UI

UI mocks

Experiment information