Release Notes¶

InQuanto 4.1.0¶

17 December 2024

Added preoptimize_passes to SparseStatevectorProtocol to address ansatz inconsistencies from v4 refactor.
Fix for PMSV due to changes in tket’s opgroup labelling.

InQuanto 4.0.1¶

28 November 2024

Improved statevector caching performance.

InQuanto 4.0.0¶

19 November 2024

General changes
- InQuanto API is now standardiZed to use American English spelling (optimiZe, diagonaliZe, etc.).
- [SymbolDict] and [SymbolSet] have been replaced with Python dictionaries. This is most noticeable in ansatz and protocols, for example to_dict and from_dict methods have been removed.
- Two new extensions: inquanto-nexus and inquanto-cutensornet
- InQuanto now supports Numpy v2+, along with other related dependency updates.
- Default numerical tolerances are now defined only in core.
Protocols have had a major refactor to enable better interaction with circuit compilation using tket passes. See e.g. compile_circuits() and the protocols manual.
- Circuit protocols must now explicitly be compiled. Therefore the circuit protocol workflow involves: instancing, building, compiling, running, and evaluation.
- build methods now creates the set of uncompiled measurement circuits.
- We recommend compile_circuits as the only place to apply circuit compression, and therefore some implicit pytket passes which were applied in (e.g.) ansatz construction are now carried to the protocol as preoptimize_passes
- Because of the above, state circuits contained in ansatz classes may contain undecomposed circuit boxes. Therefore, certain metrics of the ansatz (such as from using ansatz.state_circuit.n_2qb_gates) may not accurately reflect the final circuits in the protocol.
- SPAM calibrate() method now has an option to be performed asynchronously, to manage job delivery.
- HadamardTestDerivative has been refactored for easier computation of real, imaginary, or complex components.
- Statevector protocols have been refactored to generalize for tensor network based protocols. Use of hashing has been made more efficient.
- Modified protocols to allow Quantinuum Nexus client instances through the inquanto-nexus extension.
Operators
- Operator mappings are now performed using C++ routines.
- Added FCIDumpRestricted and FCIDumpUnrestricted for importing systems from other chemistry codes.
- Double factorization classes have been refactored.
  - Added double factorization for unrestricted-spin systems (see double_factorization_h3.py in examples) double_factorize().
  - Added ‘core tensor’ method for double factorization XDFCoreTensor
- InQuanto’s QubitOperator has been fully separated from pytket’s QubitPauliOperator (see: pytket docs).
Ansatzes
- Added unsympify() to convert from sympy to floats to avoid symbolic circuit submission. For example it will replace instances of sympy.pi with float(3.14..).
- Improved performance of ansatz hashing and comparison operations.
- to_QubitState() uses efficient implementation by default.
- State circuits contained in ansatz classes may now contain undecomposed circuit boxes.
- As such, approximate state resource estimation should be performed with circuit_resources()
- SHSV problem is solved.

InQuanto-PySCF 2.0.0¶

19 November 2024

[SymbolDict] and [SymbolSet] have been replaced with Python dictionaries.
API is now standardiZed to use American English spellings of words.
Added an example for computing oscillator strengths between VQE ground states and VQD excited states.
Performance improvements through better use of numpy.einsum and in-memory pyscf ao2mo usage.

InQuanto-NGLView 0.8.0¶

19 November 2024

API is now standardiZed to use American English spelling.
Update pint version dependency due to numpy compatibility.

InQuanto-Phayes 0.3.0¶

19 November 2024

Align dependencies with other packages.

InQuanto-Nexus 0.1.0¶

19 November 2024

Added support for the Quantinuum Nexus client for storing and managing projects, and compilation and execution jobs.
- See documentation on Nexus usage (see: InQuanto-Nexus and Quantinuum nexus ).

InQuanto-cuTensorNet 0.0.0¶

19 November 2024

Created CuTensorNetProtocol for performing state-vector and shot emulation using the pytket-cutensornet backend.
See examples.

InQuanto 3.7.0¶

22 October 2024

Added asynchronous calibration to SPAM so hardware delivery is controllable
Licensing modified for improved Quantinuum Nexus usage

InQuanto 3.6.1¶

3 September 2024

Add QCM4Computable computable for quantum computed moments up to the 4th order
Make remaining protocol classes compatible with ProtocolList by adding build_protocols_from() method definitions
New examples for the factorized overlap protocol classes (e.g. FactorizedOverlap) with comparisons to HadamardTestOverlap
Improve the information given in the dataframe outputted by protocols using the dataframe_circuit_shot() function
Add to_QubitPauliString(), from_QubitPauliString() methods to QubitOperatorString class, and to_QubitPauliOperator(), and from_QubitPauliOperator() methods to the QubitOperator class to allow the conversion of QubitOperator to/from the pytket accessible pytket.utils.QubitPauliOperator
Add all_nontrivial_qubits() property to QubitOperator, QubitOperatorString, and QubitOperatorList classes
InQuanto now throws a warning when there are dependency conflicts or required upgrades
Ensured calls to numpy.einsum correctly scale via optimize keyword argument
Make InQuanto installable with poetry
Bug fixes:
- Fixed load_xyz() defaulting to Geometry base class when called
- Fix bug where integer type parameters could cause unwarranted rounding of non-integer quantities giving incorrect results

InQuanto 3.5.8¶

23 August 2024

Hotfix release latest pytket version

InQuanto 3.5.7¶

2 August 2024

Hotfix pytket v1.31 dependency issues

InQuanto 3.5.6¶

10 July 2024

Hotfix sympy v1.13 dependency issues

InQuanto 3.5.5¶

4 June 2024

Upgraded Zeus dependency to v0.0.1b6

InQuanto 3.5.4¶

13 May 2024

Upgraded Sourcedefender dependency to v14.0

InQuanto 3.5.3¶

10 May 2024

Upgraded Sourcedefender dependency to v13.0

InQuanto 3.5.2¶

10 May 2024

Upgraded Sourcedefender dependency to v12.0

InQuanto 3.5.1¶

07 May 2024

Resolved scipy dependency conflict.
Bug fixes:
- Fixed a slowdown in qubit mappings caused in 3.5.0. Mapping speed should now be consistent with InQuanto versions <3.5.0.
- qubit_encode() for integral operators now takes the qubits argument, allowing QubitMappingBravyiKitaev support.

InQuanto 3.5.0¶

24 April 2024

Introducing new protocols: ComputeUncomputeFactorizedOverlap and SwapFactorizedOverlap for computing complex overlaps.
Expanded ProtocolList support to the following averaging protocols:
End-to-end runners for averaging protocols now support symbolic compilation with the compile_symbolic argument (see get_runner()).
Protocols now allow control over pytket optimisation_level arg at the build() step.
Updated pandas dependency to version 2.2.
Added the qubit_encode() method for FermionState and FermionStateString.
Bug fix: The complex_type arg in HadamardTestDerivative now supports string input.

InQuanto 3.4.2¶

11 April 2024

Update examples algorithm_vqe_varCI.py and algorithm_vqe_varCI_h3p.py to use up-to-date MultiConfigurationAnsatz API
Bug fixes:
- Corrected Qiskit API usage in get_noisy_backend()
- PMSV input symmetry validation is now more robust
- Fixed symbol_substitution() in FermionSpaceAnsatzChemicallyAwareUCCSD

InQuanto 3.4.1¶

08 March 2024

Upgrade Sourcedefender dependency

InQuanto 3.4.0¶

07 March 2024

Significant improvements to circuit depth in MultiConfigurationState and MultiConfigurationAnsatz

New ansatz MultiConfigurationStateBox, uses pytket’s StatePreparationBox

New composite computable ExpectationValueSumComputable

New helper methods in inquanto.ansatzes: rotate_ansatz_restricted() and reference_circuit_builder()

New examples for MultiConfigurationStateBox , inquanto.operators.ChemistryRestrictedIntegralOperator.double_factorize(), and using CCSD amplitudes as VQE initial parameters.

Note: examples are now split between inquanto examples and inquanto-extension examples

disp (display) option added for MinimizerSPSA

Major improvements to API documentation throughout

Bug fixes:

Fixed ‘=’ operator for SymbolDict

Several fixes in AlgorithmSCEOM

symbol_substitution() with string maps now works for all ansatzes

InQuanto 3.3.1¶

31 January 2024

Bug fix relating to non-symbolic computables in PauliAveraging protocol runner

InQuanto 3.3.0¶

30 January 2024

Support for Python 3.12. Drop support for Python 3.9
New method dataframe_partitioning() in PauliAveraging protocol for seeing relation between Pauli words and circuits.
Driver for generating transverse-field Ising model hamiltonians (for example DriverIsing1D)
Introducing ProtocolList. Basic support for collecting PauliAveraging protocols with different states.
Bug fix relating to qubit indexing error in symmetry analysis of QubitOperators
Other bug fixes and development improvements

InQuanto-PySCF 1.5.0¶

30 January 2024

get_double_factorized_system() method added for PySCF drivers
Python 3.12 support and deprecate 3.9

InQuanto-NGLView v0.7.1¶

30 January 2024

Python 3.12 support and deprecate 3.9
Hotfix dependency update qcelemental

InQuanto-Phayes v0.2.0¶

30 January 2024

Python 3.12 support and deprecate 3.9

InQuanto 3.2.1¶

19 January 2024

Bug fix relating to symbol substitution in PauliAveraging protocol runner

InQuanto 3.2.0¶

12 January 2024

Three new computables added
- OverlapMatrixComputable for representing the general overlap matrix between two states \(S_{ij} = <\Psi_i|\hat{O}|\Psi_j>\)
- NonOrthogonalMatricesComputable for representing matrices used by the non-orthogonal quantum eigensolver method
- SCEOMMatrixComputable for representing the Quantum Self Consistent Equation of Motion matrix

New protocol added ProjectiveMeasurements for measuring the probabilities of the basis states
Added the simultaneous perturbation stochastic approximation (SPSA) minimizer MinimizerSPSA
Added trotterize_as_linear_combination() to support second order
pd_safe_eigh() utility method now available to solve generalized eigenvalue problem \(HC = SCE\)
Express files added for Purvis’ BeH2 potential energy surface (see the express manual page)
Added method to report point_group from a FermionSpace point_group()
Added Cholesky decomposition option to double factorization
Added ensure_hermitian() to truncate non-hermitian matrix elements
Minor fix to DMET typing
Added launching and retrieving methods to QPE algorithms and protocols
Improved state hashing and comparison methods
Allow Pauli partitioning in HadamardTestOverlap to reduce measurement circuits (direct option)
Improved symbol substitution (parameter) errors and flexibility
Various other minor bug-fixes and improvements

InQuanto-PySCF 1.4.0¶

14 December 2023

Functions to get RDMs and PDMs from CASCI wavefunctions (e.g. get_casci_12rdms() )
Bug fixes

InQuanto 3.1.2¶

28 November 2023

Hotfix for PMSV by adding stabilizer tolerances
Support for error mitigation when using protocol.build_from()
Clarified the qubit_operator.eigenspectrum method

InQuanto 3.1.1¶

15 November 2023

Hotfix: added unitary check options to the DMETRHF embedding class

InQuanto 3.1.0¶

14 November 2023

Python 3.11 support
GraphColourMethod https://docs.quantinuum.com/tket/api-docs/partition.html#pytket._tket.partition.GraphColourMethod is now configurable in PauliAveraging protocol
New protocol HadamardTestOverlap for computing overlaps
Various big fixes and small improvements
Added compilation pass options to ansatz’s to_CircuitAnsatz method

InQuanto 3.0.2¶

17 October 2023

Added support for pytket 1.21
Updated GeometryMolecular to accept xyz strings
Fixes for FCIDump reader
Added exact diagonalization helper method to QubitOperators eigenspectrum()

InQuanto 3.0.1¶

10 October 2023

Improved license activation workflow
Improved API documentation

InQuanto 3.0.0¶

19 September 2023

Quantum phase estimation added
- New algorithm class AlgorithmDeterministicQPE
- New protocols for phase estimation:
Green’s functions added
- Constructed and run through the use of new composite computables:
Added Qermit integration for quantum error mitigation: https://github.com/CQCL/Qermit
Additions to the inquanto.express module
- get_system() : for quickly retrieving only the hamiltonian, fermion space, and fermion state from an express datafile
- save_h5_system() : for saving a hamiltonian, fermion space, and fermion state to h5 file
- get_noisy_backend() : for generating a simple, configurable, noisy quantum simulator
Overhaul of the inquanto.computables module
- Computables no longer have responsibility for building or running circuits. They contain ingredients for computing an observable, and relations to other computables in an expression tree. Many have been renamed, and are now categorized into three submodules (* marks a new computable, and [old name if applicable]):
- inquanto.computables.primitive (primitive objects for building computable expression trees):
  - ComputableNode *
  - ComputableSingleChild *
  - ComputableFunction *
  - ComputableList
  - ComputableTuple [Computables]
  - ComputableNDArray [ComputableArray]
- inquanto.computables.atomic (these interact directly with inquanto protocols):
  - ExpectationValue
  - ExpectationValueNonHermitian
  - ExpectationValueDerivative *
  - ExpectationValueBraDerivativeReal
  - ExpectationValueBraDerivativeImag
  - ExpectationValueKetDerivativeReal *
  - ExpectationValueKetDerivativeImag *
  - MetricTensorReal [ComputableMetricTensorReal] (no longer multiplied by factor of 1/4)
  - MetricTensorImag *
  - Overlap
  - OverlapReal *
  - OverlapImag *
  - OverlapSquared
- inquanto.computables.composite (these are composed of atomic computables, and always have “Computable” appended to the end of the name):
  - RestrictedOneBodyRDMComputable [ComputableRestrictedOneBodyRDM]
  - UnrestrictedOneBodyRDMComputable [ComputableUnrestrictedOneBodyRDM]
  - RestrictedOneBodyRDMRealComputable [ComputableRestrictedOneBodyRDMReal]
  - UnrestrictedOneBodyRDMRealComputable [ComputableUnrestrictedOneBodyRDMReal]
  - SpinlessNBodyRDMArrayRealComputable [ComputableSpinlessNBodyRDMTensorReal]
  - SpinlessNBodyPDMArrayRealComputable [ComputableSpinlessNBodyPDMTensorReal]
  - RDM1234RealComputable [ComputableRDM1234Real]
  - PDM1234RealComputable [ComputablePDM1234Real]
  - CommutatorComputable [ComputableCommutator]
  - QSEMatricesComputable [ComputableQSEMatrices]
- The following computables are no longer available:
  - ExpectationValueDerivativeReal
  - OverlapSquaredKetDerivative
Overhaul of the inquanto.protocols module.
- Protocols now have complete responsibility for building and running observable measurement circuits. They’ve also been renamed and refactored to be simpler to use. Many are pickle-able for easy pause-resume of experiments.
- New protocol names [old name]:
  - PauliAveraging [ProtocolDirect]
  - HadamardTest [ProtocolIndirect]
  - SwapTest [ProtocolCSP]
  - DestructiveSwapTest [ProtocolDSP]
  - ComputeUncompute [ProtocolVacuum]
  - PhaseShift [ProtocolPhaseShift]
  - HadamardTestDerivative [ProtocolHadamardDirectPauliY, ProtocolHadamardIndirectPauliY, ProtocolHadamardDirectPauliZ, ProtocolHadamardIndirectPauliZ]
  - HadamardTestDerivativeOverlap [ProtocolHadamardDerivativeOverlap]
  - SparseStatevectorProtocol [ProtocolStateVectorSparse]
  - BackendStatevectorProtocol [ProtocolStateVectorBackendSupport]
  - SymbolicProtocol [ProtocolSymbolic]
- The following protocols are no longer available:
  - ProtocolMidMeasurementGradient
  - ProtocolVacuumPhaseShift
inquanto.protocols.supporters submodule is no longer available.
- Error mitigation is performed using Qermit or with new noise mitigation classes (available from the inquanto.protocols module):
Changes to the inquanto.ansatzes module
- Renaming of some ansatzes [old name]:
  - MultiConfigurationAnsatz [GivensAnsatz]
  - MultiConfigurationState [MultiReferenceState]
- RealBasisRotationAnsatz is replaced by three new classes, which have improved circuit depths:
QubitState/FermionState have been restructured to make their usage more consistent with QubitOperator/FermionOperator
Many bug fixes

InQuanto-PySCF 1.3.0¶

Upgraded to PySCF 2.3.0
Many more configurable options for drivers
More memory-efficient DMET calculations with inquanto-pyscf integral operators
Support for starting calculations from checkpoint files
CAS-AC0 calculations via pyscf-ac0 https://github.com/CQCL/pyscf-ac0
WFT-in-DFT embedding, for molecular and periodic systems
Compatibility with InQuanto 3.0

2.1.1¶

27 April 2023

Some performance improvements
Fixed networkx dependence
Fixed numpy errors due to _typing
Other small fixes

A PDF version of InQuanto Version 2 documentation is also available: InQuanto-v211-PDF.

2.1.0¶

27 March 2023

Added double factorization inquanto.operators.DoubleFactorizedHamiltonian
Added Real Basis Rotation Ansatz inquanto.ansatzes.RealBasisRotationAnsatz
Addition of some C++ components and functionality (speed up compact integrals)
Adding various tools for manipulating inquanto.operators.QubitOperatorList
Qubitwise commutativity + improved tools for general commutativity
Added inquanto.operators.OperatorList.sublist() and some inquanto.algorithms.adapt.AlgorithmFermionicAdaptVQE convenience methods.
Hotfix to inquanto.geometries.GeometryMolecular.save_xyz()
Extra inquanto.algorithms.time_evolution.AlgorithmVQS examples
Upgrading Python support (up to 3.11) in line with pytket
Remove restriction on pytket-qiskit dependency requirement
inquanto.embeddings.dmet docstrings
Improved Impurity DMET exact diagonalization and VQE solver
Improved API documentation
inquanto-pyscf : inquanto.extensions.pyscf.fmo.FMO, NEVPT2 methods e.g. get_nevpt2_correction(), and fixes.
inquanto-nglview: improved API docs and fixes

2.0.0¶

5 December 2022

Re-designed ansatzes to provide more uniform interface, allow multireference calculations and easier custom ansatz development.
New ansatz classes added. CircuitAnsatz, TrotterAnsatz, GivensAnsatz, MultiReferenceState, LayeredAnsatz..
ProtocolStateVectorSparse now caches results.
New ProtocolSymbolic class allows for evaluation of expressions with symbolic ansatzes.
Implementation of the QRDM-NEVPT2 method. Several new computables for RDM calculations.
Add support for real-time evolution in AlgorithmVQS.
New CompactTwoBodyIntegralsS8 for eight-fold symmetric compact integrals.
Computable.cost_estimate and cost_estimate_elementwise for circuit cost estimation on Quantinuum hardware.
kupCCGSpD refactored to kUpCCGSDSinglet.
Pyscf refactored to PySCF.

1.3.0¶

22 November 2022

Fix FermionSpace.construct_number_alpha_operator and FermionSpace.construct_number_beta_operator.
Add FermionOperator.to_latex and QubitOperator.to_latex.
Add descriptions to files listed by inquanto.express.list_h5.
Fix bug where the PMSV supporter was sometimes not applied.
Add QubitOperatorList.all_qubits and QubitOperatorList.to_sparse_matrices.
Allow ChemistryRestrictedIntegralOperator and FermionOperator to be constructed from FCI files.
Add QubitState.from_ndarray.
Remove QubitSpace.generate_qubit_trotter_operator.
Add exponentiation and Trotterization functionality to QubitOperator and QubitOperatorList.

1.2.2¶

22 September 2022

Fixed a typo in auxiliary expressions in ADAPT and VQS.

1.2.1¶

21 September 2022

Added import for QubitMappingParity at module level.

1.2.0¶

16 September 2022

Added optional compiler_passes argument to Computable.run().
Logger configured.
Workflow to synchronize branches develop and research-develop improved.
Compact 2-body integrals implemented.
ChemistryUnrestrictedIntegralOperator._freeze() implemented.
Ansatzes deep copy implemented.
HVA fixed.
Convenience method to convert FermionOperator to integral operators implemented.
Imaginary gradients sign fixed in protocols.
Various FermionOperator helper methods implemented (is_hermitian, is_antihermitian, is_self_inverse etc.).
Disable QubitOperatorString to be constructed with two Paulis on the same qubit.
save_h5 and load_h5 implemented for spaces classes.
Added custom_prefix for circuit names to Computable.run(), .launch() and .generate_circuits().
Improvements to API documentation and type hints: Ansatzes, Algorithms, Mappings, Core, Geometries, Operators, Symmetry, Embeddings.
Fix exponent order issue in FermionSpaceStateExpChemicallyAware.

1.1.0¶

29 July 2022

toeplitz_decomposition added for QubitOperator.
ProtocolSymbolic added for symbolic evaluation of expectation values and computables.
Fixed a bug in Gradients (State vector protocols, ProtocolPhaseShift() and ProtocolHadamardDirectPauliY())
Fixed triplet excitations generator (can be used in QSE now)
General improvements in express module
- MetricTensor can now be used as part of Computables objects
Padding methods added to QubitOperator
Improved documentation

1.0.5¶

04 July 2022

Added ability for authentication by password.
Enabled support for any pytket v1.x release.

1.0.4¶

15 June 2022

Various minor bugfixes.
Improved validation output.
Updated sourcedefender dependency version.

1.0.3¶

10 June 2022

Added additional tests for parameter classes.
Fix for type safety in trigonometric methods.
Allowed api-key to be fetched from keyring or environment variable.

1.0.2¶

01 June 2022

Fixed bug in contracted systems symmetries.
Added finite difference metric tensor tests.

1.0.1¶

25 May 2022

Fixed bug in contracted systems symmetries.
Fixed bug in AlgorithmVQE property.
Fixed time-limit issue on decryption.