v1.10.0

Enhancements

• tensor network fitting: add method="tree" for when ansatz is a tree - tensor_network_fit_tree
• tensor network fitting: fix method="als" for complex dtype networks
• tensor network fitting: allow method="als" to use a iterative solver suited to much larger tensors, by default a custom conjugate gradient implementation.
• tensor_network_distance and fitting: support hyper indices explicitly via output_inds kwarg
• add tn.make_overlap and tn.overlap for computing the overlap between two tensor networks, $\langle O |T \rangle$, with explicit handling of outer indices to address hyper networks. Add output_inds to tn.norm and tn.make_norm also, as well as the squared kwarg.
• replace all numba based paralellism (prange and parallel vectorize) with explicit thread pool based parallelism. Should be more reliable and no need to set NUMBA_NUM_THREADS anymore. Remove env var QUIMB_NUMBA_PAR.
• Circuit: add dtype and convert_eager options. dtype specifies what the computation should be performed in. convert_eager specifies whether to apply this (and any to_backend calls) as soon as gates are applied (the default for MPS circuit simulation) or just prior to contraction (the default for exact contraction simulation).
• tn.full_simplify: add check_zero (by default set of "auto") option which explicitly checks for zero tensor norms when equalizing norms to avoid log10(norm) resulting in -inf or nan. Since it creates a data dependency that breaks e.g. jax tracing, it is optional.
• schematic.Drawing: add shorten kwarg to line drawing and curve drawing and examples to the docs.
• TensorNetwork: add .backend and .dtype_name properties.

PRs:

• Circuit: add default dtype and convert_eager options by @jcmgray in #273
• add fit(method="tree") and fix ALS for complex TNs by @jcmgray in #274

Full Changelog: v1.9.0...v1.10.0

v1.9.0

Breaking Changes

• renamed MatrixProductState.partial_trace and MatrixProductState.ptr to MatrixProductState.partial_trace_to_mpo to avoid confusion with other partial_trace methods that usually produce a dense matrix.

Enhancements:

• add Circuit.sample_gate_by_gate and related methods CircuitMPS.reordered_gates_dfs_clustered and CircuitMPS.get_qubit_distances for sampling a circuit using the 'gate by gate' method introduced in https://arxiv.org/abs/2112.08499.
• add Circuit.draw for drawing a very simple circuit schematic.
• Circuit: by default turn on simplify_equalize_norms and use a group_size=10 for sampling. This should result in faster and more stable sampling.
• Circuit: use numpy.random.default_rng for random number generation.
• add qtn.circ_a2a_rand for generating random all-to-all circuits.
• expose qtn.edge_coloring as top level function and allow layers to be returned grouped.
• add docstring for tn.contract_compressed and by default pick up important settings from the supplied contraction path optimizer (max_bond and compress_late)
• add Tensor.rand_reduce for randomly removing a tensor index by contracting a random vector into it. One can also supply the value "r" to isel selectors to use this.
• add fit-zipup and fit-projector shorthand methods to the general 1d tensor network compression function
• add MatrixProductState.compute_local_expectation for computing many local expectations for a MPS at once, to match the interface for this method elsewhere. These can either be computed via canonicalization (method="canonical"), or via explicit left and right environment contraction (method="envs")
• specialize CircuitMPS.local_expectation to make use of the MPS form.
• add PEPS.product_state for constructing a PEPS representing a product state.
• add PEPS.vacuum for constructing a PEPS representing the vacuum state $|000\ldots0\rangle$.
• add PEPS.zeros for constructing a PEPS whose entries are all zero.
• tn.gauge_all_simple: improve scheduling and add damping and touched_tids options.
• qtn.SimpleUpdateGen: add gauge difference update checking and tol and equilibrate settings. Update .plot() method. Default to a small cutoff.
• add psi.sample_configuration_cluster for sampling a tensor network using the simple update or cluster style environment approximation.
• add the new example doc Exploring circuit sampling with various methods

v1.8.4

What's Changed

• fix MPS sample handling of RNG seed by @kevinsung in #248
• fix bug in applying MPO lazily to MPS (#246)

New Contributors

• @kevinsung made their first contribution in #248

Full Changelog: v1.8.3...v1.8.4

v1.8.3

Enhancements:

• support for numpy v2.0 and scipy v1.14
• add MPS sampling: MatrixProductState.sample_configuration and MatrixProductState.sample (generating multiple samples) and use these for CircuitMPS.sample and CircuitPermMPS.sample.
• add basic plot method for [SimpleUpdate] classes
• add edges_1d_chain for generating 1D chain edges
• operatorbuilder: better coefficient placement for long range / non-local MPO building

Full Changelog: v1.8.2...v1.8.3

v1.8.2

Enhancements:

• TNOptimizer can now accept an arbitrary pytree (nested combination of dicts, lists, tuples, etc. with TensorNetwork, Tensor or raw array_like objects as the leaves) as the target object to optimize.
• TNOptimizer can now directly optimize Circuit objects, returning a new optimized circuit with updated parameters.
• Circuit: add .copy(), .get_params() and .set_params() interface methods.
• Update generic TN optimizer docs.
• add tn.gen_inds_loops for generating all loops of indices in a TN.
• add tn.gen_inds_connected for generating all connected sets of indices in a TN.
• make SVD fallback error catching more generic (by @mlxd in #238)
• fix some windows + numba CI issues.
• approx_spectral_function add plotting and tracking
• add dispatching to various tensor primitives to allow overriding

New Contributors

• @mlxd made their first contribution in #238

Full Changelog: v1.8.1...v1.8.2

v1.8.1

Enhancements:

• CircuitMPS now supports multi qubit gates, including arbitrary multi-controls (which are treated in a low-rank manner), and faster simulation via better orthogonality center tracking.
• add CircuitPermMPS, more docs here: https://quimb.readthedocs.io/en/latest/tensor-circuit-mps.html
• add MatrixProductState.gate_nonlocal for applying a gate, supplied as a raw matrix, to a non-local and arbitrary number of sites. The kwarg contract="nonlocal" can be used to force this method, or the new option "auto-mps" will select this method if the gate is non-local (#230)
• add MatrixProductState.gate_with_mpo for applying an MPO to an MPS, and immediately compressing back to MPS form using tensor_network_1d_compress
• add MatrixProductState.gate_with_submpo for applying an MPO acting only of a subset of sites to an MPS
• add MatrixProductOperator.from_dense for constructing MPOs from dense matrices, including an only subset of sites
• add MatrixProductOperator.fill_empty_sites for 'completing' an MPO which only has tensors on a subset of sites with (by default) identities
• MatrixProductState and MatrixProductOperator, now support the sites kwarg in common constructors, enabling the TN to act on a subset of the full L sites.
• add TensorNetwork.drape_bond_between for 'draping' an existing bond between two tensors through a third
• add Tensor.new_ind_pair_with_identity
• TN2D, TN3D and arbitrary geom classical partition function builders (TN_classical_partition_function_from_edges) now all support outputs= kwarg specifying non-marginalized variables
• add simple dense 1-norm belief propagation algorithm D1BP
• add qtn.enforce_1d_like for checking whether a tensor network is 1D-like, including automatically adding strings of identities between non-local bonds, expanding applicability of tensor_network_1d_compress
• add MatrixProductState.canonicalize as (by default non-inplace) version of canonize, to follow the pattern of other tensor network methods. canonize is now an alias for canonicalize_ [note trailing underscore].
• add MatrixProductState.left_canonicalize as (by default non-inplace) version of left_canonize, to follow the pattern of other tensor network methods. left_canonize is now an alias for left_canonicalize_ [note trailing underscore].
• add MatrixProductState.right_canonicalize as (by default non-inplace) version of right_canonize, to follow the pattern of other tensor network methods. right_canonize is now an alias for right_canonicalize_ [note trailing underscore].

Bug fixes:

• Circuit.apply_gate_raw: fix kwarg bug (#226) by @juliendrapeau
• fix for retrieving opt_einsum.PathInfo for single scalar contraction (#231).

New Contributors

• @juliendrapeau made their first contribution in #226

Full Changelog: v1.8.0...v1.8.1

v1.8.0

Breaking Changes

• all singular value renormalization is turned off by default
• TensorNetwork.compress_all now defaults to using some local gauging

Enhancements:

• add quimb.tensor.tensor_1d_compress.py with functions for compressing generic 1D tensor networks (with arbitrary local structure) using various methods. The methods are:

  • The 'direct' method: tensor_network_1d_compress_direct
  • The 'dm' (density matrix) method: tensor_network_1d_compress_dm
  • The 'zipup' method: tensor_network_1d_compress_zipup
  • The 'zipup-first' method: tensor_network_1d_compress_zipup_first
  • The 1 and 2 site 'fit' or sweeping method: tensor_network_1d_compress_fit
  • ... and some more niche methods for debugging and testing.

  And can be accessed via the unified function tensor_network_1d_compress. Boundary contraction in 2D can now utilize any of these methods.

• add quimb.tensor.tensor_arbgeom_compress.py with functions for compressing arbitrary geometry tensor networks using various methods. The methods are:

  • The 'local-early' method: tensor_network_ag_compress_local_early
  • The 'local-late' method: tensor_network_ag_compress_local_late
  • The 'projector' method: tensor_network_ag_compress_projector
  • The 'superorthogonal' method: tensor_network_ag_compress_superorthogonal
  • The 'l2bp' method: tensor_network_ag_compress_l2bp

  And can be accessed via the unified function tensor_network_ag_compress. 1D compression can also fall back to these methods.

• support PBC in tn2d.contract_hotrg, tn2d.contract_ctmrg, tn3d.contract_hotrg and the new function tn3d.contract_ctmrg.

• support PBC in gen_2d_bonds and gen_3d_bonds, with cyclic kwarg.

• support PBC in TN2D_rand_hidden_loop and TN3D_rand_hidden_loop, with cyclic kwarg.

• support PBC in the various base PEPS and PEPO construction methods.

• add tensor_network_apply_op_op for applying 'operator' TNs to 'operator' TNs.

• tweak tensor_network_apply_op_vec for applying 'operator' TNs to 'vector' or 'state' TNs.

• add tnvec.gate_with_op_lazy method for applying 'operator' TNs to 'vector' or 'state' TNs like $x \rightarrow A x$.

• add tnop.gate_upper_with_op_lazy method for applying 'operator' TNs to the upper indices of 'operator' TNs like $B \rightarrow A B$.

• add tnop.gate_lower_with_op_lazy method for applying 'operator' TNs to the lower indices of 'operator' TNs like $B \rightarrow B A$.

• add tnop.gate_sandwich_with_op_lazy method for applying 'operator' TNs to the upper and lower indices of 'operator' TNs like $B \rightarrow A B A^\dagger$.

• unify all TN summing routines into tensor_network_ag_sum, which allows summing any two tensor networks with matching site tags and outer indices, replacing specific MPS, MPO, PEPS, PEPO, etc. summing routines.

• add rand_symmetric_array, rand_tensor_symmetric TN2D_rand_symmetric for generating random symmetric arrays, tensors and 2D tensor networks.

Bug fixes:

• fix scipy sparse monkey patch for scipy>=1.13 (#222)
• fix autoblock bug where connected sectors were not being merged (#223)

Full Changelog: v1.7.3...v1.8.0

v1.7.3

Enhancements:

• qu.randn: support dist="rademacher".
• support dist and other randn options in various TN builders.

Bug fixes:

• restore fallback (to scipy.linalg.svd with driver='gesvd') behavior for truncated SVD with numpy backend.

Full Changelog: v1.7.2...v1.7.3

v1.7.2

Bug fixes:

• removed import of deprecated numba.generated_jit decorator.

Enhancements:

• add normalized=True option to tensor_network_distance for computing the normalized distance between tensor networks: $2 |A - B| / (|A| + |B|)$, which is useful for convergence checks. Tensor.distance_normalized and TensorNetwork.distance_normalized added as aliases.
• add TensorNetwork.cut_bond for cutting a bond index

Full Changelog: v1.7.1...v1.7.2

v1.7.1

What's Changed

Enhancements:

• add TensorNetwork.visualize_tensors for visualizing the actual data entries of an entire tensor network.

• add ham.build_mpo_propagator_trotterized for building a trotterized propagator from a local 1D hamiltonian. This also includes updates for creating 'empty' tensor networks using TensorNetwork.new, and building up gates from empty tensor networks using TensorNetwork.gate_inds_with_tn.
• add more options to Tensor.expand_ind and Tensor.new_ind: repeat tiling mode and random padding mode.
• tensor decomposition: make eigh_truncated backend agnostic.
• tensor_compress_bond: add reduced="left" and reduced="right" modes for when the pair of tensors is already in a canonical form.
• add qtn.TN2D_embedded_classical_ising_partition_function for constructing 2D (triangular) tensor networks representing all-to-all classical ising partition functions.

Bug fixes:

• fix bug in kruas_op when operator spanned multiple subsystems (#214)
• fix bug in qr_stabilized when the diagonal of R has significant imaginary parts.
• fix bug in quantum discord computation when the state was diagonal (#217)
• Fix empty lines in dimacs by @jjcmoon in #215

New Contributors

• @jjcmoon made their first contribution in #215

Full Changelog: v1.7.0...v1.7.1