Qiskit Transpiler Tools

Qiskit tools to streamline transpilation

Installation

Package not yet available through PyPI. Clone the main branch to your local repository. Then, install with poetry in editable mode by running

$ poetry install

in the root directory, for usage and/or development. Or simply include the root directory in your python search path.

Usage

At the moment, there is only one transpiler class implementing the simple pipeline: gate translation -> layout optimization -> noise aware mapping (mapomatic) -> dynamic decoupling.

Import the class:

from qiskit_transpiler_tools.transpiler import TranspilerSabreMapomaticDD

Construct a circuit qc or a list of circuits qc_list, load your backend. Apply different transpiler configurations:

transpiler_default = TranspilerSabreMapomaticDD(backend) # default options
transpiler1        = TranspilerSabreMapomaticDD(backend,
                                                optimization_level=3,
                                                num_transpilations=10) # layout optimization
transpiler2        = TranspilerSabreMapomaticDD(backend,
                                                optimization_level=3,
                                                num_transpilations=10,
                                                apply_mapomatic=True,
                                                apply_dd=True) # add noise aware transpilation and dynamic decoupling

qc_trans       = transpiler_default.transpile(qc)
qc_list_trans1 = transpiler1.transpile(qc_list)
qc_list_trans2 = transpiler2.transpile(qc_list + [qc])

Contributing

Fork the repository, use basic GitFlow and semantic-release. Further rules will be specified if the need arises.

License

qiskit_transpilation_tools was created by Mazen Ali. It is licensed under the terms of the MIT license.