Changelog

See the full changelog on GitHub: CHANGELOG.md

[Unreleased]

[v0.12.0] - 2026-05-16

New Features

  • Extended HybridSuperQubits.noise with table-level sweep functions: t1_table_from_spectral_density, tphi_1_over_f_table, tphi_cqps_table. Channel-specific spectral densities (S_capacitive, S_inductive, S_flux_bias_line, S_charge_impedance, S_one_over_f_flux, S_critical_current, S_andreev) are now public — a single source of truth shared by both the single-shot and sweep code paths.

  • The 7 get_t1_*_vs_paramvals methods, 2 get_tphi_*_vs_paramvals methods, get_tphi_CQPS_vs_paramvals, and the internal _get_t1_vs_paramvals / _get_tphi_1_over_f_vs_paramvals dispatchers on QubitBase are now thin wrappers around these functions. Phase B of the #17 refactor.

Breaking Changes

  • Reconciled single-shot vs sweep physics divergences. The single-shot t1_capacitive, t1_inductive, and t1_flux_bias_line methods (and the corresponding HybridSuperQubits.noise functions) now produce the same numerical results as their sweep counterparts. Previously they differed by factors of ~2 (capacitive) and used a different formula entirely (flux-bias-line). The sweep version was the physically correct one and is now the single source of truth.

  • Removed the total: bool keyword argument from t1_capacitive, t1_inductive, t1_flux_bias_line (both class methods and standalone functions) and from get_t1_*_vs_paramvals sweep methods. The kwarg was the knob that selected the buggy “SD(+ω) + SD(-ω)” summing in single-shot; with the physics reconciled there is no longer a switch to flip. Passing total=... will raise TypeError.

Bug Fixes

  • Capacitive single-shot T1 prefactor corrected from 8 to 16 to match the sweep path. Combined with the v0.11.0 Q_cap default fix, single-shot capacitive T1 values are now consistent with sweep values for the same parameters.

[v0.11.0] - 2026-05-16

New Features

  • Added HybridSuperQubits.noise module: standalone pure functions for the existing decoherence formulas (t1_capacitive, t1_inductive, t1_flux_bias_line, tphi_1_over_f, tphi_CQPS, plus the generic t1_from_spectral_density builder). Functions take eigenvalues, operator matrix elements, and scalar parameters — no qubit object required — so decoherence can be computed for any user-supplied Hamiltonian. First step of the functional refactor proposed in #17.

  • The corresponding methods on QubitBase are now thin wrappers that delegate to HybridSuperQubits.noise (covered by tests/integration/test_noise_parity.py).

Bug Fixes

  • Corrected the default capacitive quality factor prefactor in t1_capacitive from 1e6 to 1/3e-5 (~3.33e4), the physically correct value already in use by the get_t1_capacitive_vs_paramvals sweep. This makes the single-shot and sweep capacitive-T1 paths consistent. Behavior change: capacitive T1 values returned by qubit.t1_capacitive() (when Q_cap is not explicitly supplied) are now ~30x shorter than in v0.10.x.

[v0.10.9] - 2026-04-15

Bug Fixes

  • Added missing 4 * Ec prefactor to the scalar Berry contribution in Ferbo._berry_contribution().

[v0.10.8] - 2026-04-11

Breaking Changes

  • Renamed Ferbo basis values from "static" → "ballistic" and "adiabatic" → "Andreev" across the codebase.

  • Renamed the andreev_basis parameter to representation in Ferbo.wavefunction(). The parameter name remains andreev_basis in Ferbo.plot_wavefunction().

[v0.10.7] - 2026-04-04

Bug Fixes

  • Fixed CI benchmark workflow referencing removed full extra after dependencies were made required.

[v0.10.6] - 2026-04-04

Improvements

  • Removed poetry.lock from version control; not needed for a library package and was causing CI build failures when dependencies changed.

[v0.10.5] - 2026-04-04

Improvements

  • Made numpy, scipy, and qutip required dependencies so pip install HybridSuperQubits is self-contained. Removed unused scqubits from required deps (kept as optional extra).

  • Removed qutip.Qobj dependency from cos_phi/sin_phi in operators.py; they now return plain scipy sparse matrices.

[v0.10.4] - 2026-04-04

Bug Fixes

  • Fixed missing wavefunction normalization factor in Gatemonium.wavefunction(). Added the 1/sqrt(phase_zpf) divisor to match the correct normalization used in Ferbo and Fluxonium.

[v0.10.3] - 2026-04-04

Improvements

  • Replaced str type annotations with Literal types for all fixed-option string parameters across Andreev, Ferbo, Fluxonium, Gatemonium, QubitBase, and Resonator classes, improving IDE autocompletion and type safety.

[v0.10.2] - 2026-04-03

Breaking Changes

  • Replaced the rotate boolean in Ferbo.wavefunction() (now representation parameter, "ballistic" | "Andreev") and Ferbo.plot_wavefunction() (now andreev_basis parameter, "ballistic" | "Andreev").

New Features

  • Added scalar Berry phase correction to the Andreev potential via Ferbo._berry_contribution().

  • Added include_berry flag to Ferbo.potential() to toggle the Berry correction.

Testing

  • Added unit tests for Berry contribution (include/exclude, both ABS and EL flux groupings).

  • Added test verifying return_evecs and evals-only paths return consistent eigenvalues.

[v0.10.1] - 2026-02-13

Added

  • Added second derivative methods for Ferbo class: d2_hamiltonian_d_Gamma2, d2_hamiltonian_d_deltaGamma2, and d2_hamiltonian_d_er2

[v0.10.0] - 2026-02-13

Improvements

  • Fixing factors in the spectrum density value and refactoring

[v0.9.4] - 2025-10-31

Improvements

  • Version bump for patch release

[v0.9.3] - 2025-06-28

Improvements

  • Modernized CI/CD with updated GitHub Actions workflows

  • Migrated from flake8 + black to ruff for linting and formatting

  • Updated all type hints to use modern Python built-in types (dict, list, tuple)

  • Enhanced regression tests with physically realistic parameters

  • Fixed 136+ linting errors automatically

  • Improved pyproject.toml configuration with comprehensive tool settings

Breaking Changes

  • Minimum Python version updated to 3.10 (due to QuTiP dependency requirements)

Testing

  • Updated all regression test reference values with realistic physical parameters

  • Increased Ferbo test dimension to 200 for better accuracy

  • All 102 tests passing with new configuration

  • Added robust parameter validation and error handling

Documentation

  • Updated project guidelines for English-only requirements

  • Improved code documentation and type hints

  • Cleaned up development dependencies

Technical Details

  • Ferbo tests now use realistic parameters: El~0.1-0.2 GHz, Ec~1-5 GHz, Ej~1-10 GHz

  • Fluxonium tests use dimension=100 with physically consistent parameters

  • Simplified build process without manual SciPy pre-installation

  • Enhanced mypy configuration for gradual typing support

  • Updated Poetry lock file for dependency compatibility

[v0.9.2] - 2025-06-25

Improvements

  • Modernized CI/CD with updated GitHub Actions workflows

  • Migrated from flake8 + black to ruff for linting and formatting

  • Updated all type hints to use modern Python built-in types (dict, list, tuple)

  • Enhanced regression tests with physically realistic parameters

  • Fixed 136+ linting errors automatically

  • Improved pyproject.toml configuration with comprehensive tool settings

Testing

  • Updated all regression test reference values with realistic physical parameters

  • Increased Ferbo test dimension to 200 for better accuracy

  • All 102 tests passing with new configuration

  • Added robust parameter validation and error handling

Documentation

  • Updated project guidelines for English-only requirements

  • Improved code documentation and type hints

  • Cleaned up development dependencies

Technical Details

  • Ferbo tests now use realistic parameters: El~0.1-0.2 GHz, Ec~1-5 GHz, Ej~1-10 GHz

  • Fluxonium tests use dimension=100 with physically consistent parameters

  • Simplified build process without manual SciPy pre-installation

  • Enhanced mypy configuration for gradual typing support

[v0.9.1] - 2025-06-25

Main changes

  • Add show progress in get_matelements_vs_paramvals.

[v0.9.0] - 2025-06-25

Main changes

  • Enhanced Gatemonium class to support multiple transmission channels

  • Changed Delta parameter to have a default value of 44 (for Aluminum)

  • Reorganized Gatemonium potential calculation to handle multiple channels

  • Updated derivatives with respect to external phase to support multiple channels

[v0.8.9] - 2025-06-25

Main changes

  • Added Resonator class to simulate microwave resonators

  • Improved JJA class with enhanced error handling and parameter validation

  • Enhanced Ferbo class with optimized calculations and additional utilities

  • Added scqubits as a dependency for enhanced compatibility

  • Fixed several bugs and improved code quality

  • Updated documentation

[v0.8.4] - 2025-06-04

Main changes

  • Add version retrieval in the __init__.py.

[v0.2.0] - 2025-02-21

Main changes

  • Add methods state_to_density_matrix, ptrace and purity in operators.py.

  • Replace the Ferbo potential by one rotated by pi/2 in the Y axis.

  • Fix bugs for updating the SpectrumData.

  • Migrate project configuration to Poetry.

  • Add CHANGELOG.md.