Qiskit MPS Initializer

This package provides extra tools on Qiskit enabling you to initialize wavefunctions on a quantum computer using techniques based on matrix product states (MPS).

User guide

Installation

The easiest way to get started is using pip which installs the package and its dependencies in your project

pip install qiskit-mps-initializer

You can alternatively use uv, poetry or any other python package manager to install this package.

Usage

Once you have installed the package, you can import tools from it.

from qiskit_mps_initializer.datatypes import QuantumState

# wavefunction as an array of numbers, could use np.array as well
psi = [1, 3, 1, 2, 7, 8, 0, 1]
# number of layers of two-qubit gates to use for the initializer
number_of_mps_layers = 2

# create the state object
state = QuantumState.from_dense_data(data=psi, normalize=True)

# generate the initializer circuit for this state
circuit = state.generate_mps_initializer_circuit(number_of_layers=number_of_mps_layers)

# the circuit object is a qiskit.QuantumCircuit object which
# means you can do whatever you could natively do in Qiskit
circuit.draw('mpl')

Documentation

The project's documentation is located here. Check it out for more information.

[!CAUTION] This project is in alpha. This means you should expect drastic changes in the api in later releases.

Project credits

Tools

• Project and dependency manager: uv
• Linter: ruff
• Formatter & style: ruff
• Static typechecking: pyright (ty is currently in beta, pyrefly is also another candidate, both built using Rust)
• Unit testing: pytest (no Rust-based alternative)
  • Randomization: hypothesis
• Documentation: mkdocs
  • API auto-generation: mkdocstrings-python
  • Theme: mkdocs-material

Dependencies

• Quantum circuits: qiskit
• Data modeling & validation: pydatic
  • Numpy support: pydantic-numpy
• Tensor networks: quimb