README.md

LibMark2

A simple python package for sending and pulling VQE run results to/from WebMark2

Getting started

You can check the currently viable guide on how to get started with Libmark from the quickstart guide.

There is also an example notebook available which showcases the functionality.

Using the package

Make sure that the package is somewhere in your PYTHONPATH, then after using

from LibMark2.quantmark.tracker import get_tracker

qresult = get_tracker(optimizer, 'TOKEN')

where "optimizer" is simply a string (name of optimizer used). The token can be acquired from the Quantmark website. 'get_tracker' returns the correct Result object.

Data can be added to the object using

qresult.add_run(results, molecule, hamiltonian, ansatz)

Where 'results' is an object returned by tq.minimize.

Data can be sent to WebMark2, by calling the push function of any Result object

qresult.push()

Want to save the data?

qresult.save()

will save a JSON file to the root.

Pushing

# Generic batch optimize for "any" VQE
def run_vqe(molecules, hamiltonian_function, ansatz_function, optimizer, silent=True, **ansatz_kwargs):
    results = []
    qresult = get_tracker(optimizer, 'TOKEN_HERE') # Start quantmark tracking
    for i, molecule in enumerate(molecules):
        print(str(i+1)+"/"+str(len(molecules)), end="\t")
        print("Creating the Hamiltonian.", end="\t")
        H = hamiltonian_function(molecule)
        n_qubits = len(H.qubits)
        print("Creating ansatz.", end="\t")
        U = ansatz_function(molecule=molecule, n_qubits=n_qubits, **ansatz_kwargs)
        print("Creating objective function")
        E = tq.ExpectationValue(H=H, U=U)
        variables = {k:0.0 for k in U.extract_variables()}
        print("Optimizing.", end="\t")
        result = tq.minimize(objective=E, method=optimizer, initial_values=variables, silent=silent)
        print()
        results.append(result)
        qresult.add_run(result, molecule, H, U) # Adding the intermediate result
    print("Done")
    qresult.push() # Push results to the server
    qresult.save() # Save the data to a JSON file
    return results

Pulling

You can pull any public result (or your own result) from WebMark:

from LibMark2.quantmark.api import get_experiment, get_data

data = get_data(id, 'TOKEN_HERE') # Download all available information as a dict
experiment = get_experiment(id, 'TOKEN_HERE') # Download experiment
results = experiment.run_experiment() # Run the experiment

"results" will contain a list of tuples (distance, results)