Quantum Inspire

Welcome to the repository for the Quantum Inspire tool. The goal of this project to offer basic support for interacting with the Quantum Inspire platform. For example, you can manage your projects, algorithms and jobs on the Quantum Inspire system from your command line, or using the API directly in your Python code.

The tool is required for logging in to the Quantum Inspire systems. For example, if you would like to use the QI Qiskit or Pennylane plugins.

Installation

The recommended way of installing the CLI is to use pipx. (Conda users should refer to the subsection below). After following the pipx installation instructions, run the following command:

pipx install quantuminspire

Afterwards, running qi --help should show a help menu.

Installation using Conda

Check you Conda version with conda --version, then run the appropriate command:

Conda >= 26.3:

conda create --file https://conda.quantum-inspire.com/cli.yml

Conda < 26.3:

conda env create --file https://conda.quantum-inspire.com/cli.yml

By default, the environment that will be created will be named quantuminspire; use the -n/--name <CUSTOM-ENV-NAME> option to overwrite this.

After creating the environment, you need to activate it (use your custom environment name if you have defined it):

conda activate quantuminspire

Using the CLI to login

In order to login to the platform, run the following command:

qi login

This will open a browser window that will allow you to login or create an account. By default this command will login to the production environment, but the command accepts an argument for a different host URL if needed (e.g. for beta testing purposes).

Logout

To log out from the platform, run:

qi logout

If no hostname is specified, the default host is used.

Run a job

To run a job on Quantum Inspire without persisting any settings, provide a file and a backend type ID:

qi run --file <path to .cq or .py file> --backend-type-id <int>

This will return a job ID. You can optionally specify --num-shots, --store-raw-data, and --name. Use --persist together with --name to store the project and algorithm settings locally for repeated use. If you use the --persist flag, you first have to initialize a project with qi projects init <name>.

IMPORTANT: Submitting a (python) .py file for hybrid algorithms requires a specific format. Not all .py files are valid for submission.

Projects

Initialize a project

Create a new project on the platform and store its settings locally in the current working directory (project root):

qi projects init <name> [description]

Examples

Initialize a project with only a name:

qi projects init "My Project"

Initialize a project with a name and description:

qi projects init "My Project" "Optional description"

List projects

List all projects belonging to the current user. You can optionally filter projects by name, either using substring matching (default) or exact matching.

Usage

qi projects list

Options

• --quiet, -q Only print the IDs of the projects.

• --name <name> Filter projects by name or description using substring matching.

• --exact When used with --name, only return projects whose name exactly matches the given value.

Examples

List all projects:

qi projects list

List projects whose name or description contains "quantum":

qi projects list --name quantum

List projects whose name is exactly "my-project":

qi projects list --name my-project --exact

Print only the project IDs:

qi projects list --quiet

Delete projects

Delete one or more projects by ID, by name, or delete all projects after confirmation.

⚠️ Deleting projects is irreversible. You will always be prompted for confirmation before deletion.

Usage

qi projects delete [PROJECT_IDS...]

Options

• PROJECT_IDS One or more project IDs to delete.

• --name <name> Delete projects matching this name or description pattern.

• --exact When used with --name, only delete projects whose name exactly matches the given value.

Behavior and precedence

• If project IDs are provided, --name and --exact are ignored.

• If no IDs are provided:

  • --name filters projects by substring match
  • --name + --exact performs an exact name match

• If no IDs and no name are provided, all projects will be deleted after confirmation.

Examples

Delete projects by ID:

qi projects delete 1 2 3

Delete projects whose name or description contains "quantum":

qi projects delete --name quantum

Delete the project with the exact name "my-project":

qi projects delete --name my-project --exact

Use list and delete together to delete projects by name:

qi projects delete $(qi projects list --name quantum --quiet)

Delete all projects:

qi projects delete

Algorithms

Algorithms allow you to persist settings (file path, backend type, number of shots, raw-data flag) locally for repeated job submission. Algorithms must belong to a project initialized with qi projects init. You can get the status, results and final result using the algorithm name. This will retrieve the specified resource for the latest job that was run using the algorithm. If you want to retrieve a resource for a specific job, use the qi jobs commands instead.

Initialize an algorithm

qi algorithms init <name> <path> <backend_type_id>

Get algorithm job status

qi algorithms status <algorithm_name>

Use --wait and --timeout <seconds> to wait for the job to complete.

Get algorithm job logs

qi algorithms logs <algorithm_name>

Use --n-logs to return early if you already know how many logs you expected, --poll-interval <seconds> to set the polling interval (default: 5s) and --timeout <seconds> to set the timeout (default: 30s).

Get algorithm results

qi algorithms results <algorithm_name>

Use --save to save the results to a JSON file in the current directory.

Get algorithm final result

qi algorithms final-result <algorithm_name>

Use --save to save the final result to a JSON file in the current directory.

Note: This object will always be generated. In the case of a quantum circuit, the result and final result will be the same. For hybrid algorithms, the final result is a free form datastructure that could for example be used for the aggregation of data. This is generated in the finalize step.

Jobs

Jobs are created when you run qi run. Use the job ID printed by that command to check status and retrieve results. You can also get the status, results and final result using the algorithm name with the qi algorithms commands. This will retrieve the specified resource for the latest job that was run using the algorithm.

Inspect a job

Show all properties of a specific job:

qi jobs inspect <job_id>

Get job status

qi jobs status <job_id>

Use --wait to wait for the job to complete before returning, and --timeout <seconds> to set the timeout (default: 60).

Get job logs

qi jobs logs <job_id>

Use --n-logs to return early if you already know how many logs you expected, --poll-interval <seconds> to set the polling interval (default: 5s) and --timeout <seconds> to set the timeout (default: 30s).

Get job results

qi jobs results <job_id>

Use --save to save the results to a JSON file in the current directory.

Get job final result

A job also always contains a final result. This object can be queried with the following command.

qi jobs final-result <job_id>

Use --save to save the final result to a JSON file in the current directory.

Note: This object will always be generated. In the case of a quantum circuit, the result and final result will be the same. For hybrid algorithms, the final result is a free form datastructure that could for example be used for the aggregation of data. This is generated in the finalize step.

Backends

The list of backends (and some of their properties) can be retrieved using the following command, from which the id field can be read.

qi backends list

To see all properties of a specific backend, use:

qi backends get <backend-type-id>

Compile files

The CLI can be used to compile cQASM (.cq) files against a specific backend type.

IMPORTANT: Hybrid Python (.py) files are not supported for compilation.

qi files compile --file <filename> --backend-type-id <int>

You can optionally specify a --compile-stage to control up to which stage the algorithm is compiled (default: Decomposition). Alternatively, provide --name to use the settings of a previously initialized algorithm.

Config

Get a configuration value

qi config get <key>

Use --algorithm <name> to retrieve a setting for a specific algorithm.

Set a configuration value

qi config set <key> <value>

Use --algorithm <name> to set a setting for a specific algorithm, or --user to set a global user-level setting (e.g. default_host).