Add lattice-utils to dwave-experimental by SebastianGitt · Pull Request #44 · dwavesystems/dwave-experimental

SebastianGitt · 2026-03-13T20:49:49Z

Add a minimal version of LatQA to dwave-experimental as a submodule named lattice-utils. Currently supports examples of a 1D Ising Chain and a 2D dimerized triangular lattice. Tests still need to be added to the pull request.

-move initialization of num_spins to subclass

thisac

Thanks @SebastianGitt! First pass of the lattice submodule.

I'd recommend running a formatter (black) since all of these files are new and there are a few places with incorrect spacing, etc.
I think the saving/loading of embeddings can be refined a bit, but I'll have a closer look at it later.

thisac · 2026-04-14T01:16:45Z

+# from latqa.experiment import *
+# from latqa.lattice import *
+# from latqa.observable import *
+# from latqa.analysis import *
+# from latqa._paths import *


Forgotten? Should probably just be

import dwave.experimental.lattice_utils.experiment import dwave.experimental.lattice_utils.lattice ...

unless there any any general functions that should be accessible directly from the dwave.experimental.lattice_utils namespace.

thisac · 2026-04-14T01:19:14Z

+    seed: int | None = None,
+    skipnan: bool = True,
+) -> list[float]:
+    """Compute bootstrap estimates of a statistic."""


Expand on docstring, add args, return, etc.

thisac · 2026-04-14T01:19:21Z

+    repetitions: int,
+    seed: int | None = None,
+) -> Iterator[NDArray]:
+    """Generate resampled indices."""


thisac · 2026-04-14T01:19:29Z

+
+
+def confidence_interval(array: NDArray, width: float = 0.95) -> tuple[float, float, float]:
+    """Ravel and take the quantiles; return median and error bar lengths."""


thisac · 2026-04-14T18:29:49Z

+def generate_bootstrap_indices(
+    size: int,
+    repetitions: int,
+    seed: int | None = None,


Better to use NumPy's random.default_rng instead (see note on random.seed doc).

Suggested change

seed: int | None = None,

rng: np.random.Generator | None = None,

and then use rng.choice(...) below.

thisac · 2026-04-24T21:08:53Z

+        u: Hashable,
+        v: Hashable | None = None,
+    ) -> Iterator[tuple[int, int]]:
+        """Should also work for chains!  These can be thought of as self-loops."""


Update docstring and add args, returns.

thisac · 2026-04-24T21:10:20Z

+        if u == v or v is None:
+            # Interior chain connectivity.
+            # Generic version: add all possible edges.
+            return ((0, 1),)


Why return tuple[tuple[int, int]] instead of just tuple[int, int]?

It represents a tuple of edges. In this case it's only a single edge but for other geometries it's expected that get_chain_connectivity() returns multiple edges

thisac · 2026-04-24T21:33:27Z

+import os
+from pathlib import Path
+from collections.abc import Hashable
+from numbers import Integral


Not used.

Suggested change

from numbers import Integral

thisac · 2026-04-24T21:33:45Z

+__all__ = ['Lattice']
+


Suggested change

__all__ = ['Lattice']

__all__ = ['Lattice']

thisac · 2026-04-24T21:37:39Z

+        if not hasattr(self, "num_spins"):
+            raise AttributeError(f"{type(self).__name__} subclass must initialize self.num_spins")


This would make it pointless to instantiate Lattice directly, no? Why not make it into an abstract base-class instead, adding in potentially relevant abstract methods, etc.?

-added more comprehensive docstrings -fixed formatting issues

…dwave-experimental into add-lattice-utils

…utils

-removed spin reversal transform functionality -uses automorphism module already in dwave-experimental -uses data classes for experiment configs

-removed spin reversal transform functionality -uses automorphism module already in dwave-experimental -uses data classes for experiment configs -formatted using black

andrew-d-king

Looks generally good; I am not able to pore over it in detail but it all seems sensible.

andrew-d-king · 2026-05-15T13:06:32Z

+        response_dict = {}
+        call_dict = {}
+
+        for index, param in enumerate(parameter_list):


I would like to see the progress bars restored here.

andrew-d-king · 2026-05-15T13:09:03Z

+        ret = parameter_list.copy()
+        for entry in ret:
+            if "anneal_time" in entry:
+                entry["anneal_time"] = np.round(entry["anneal_time"], 6)


Always nice to see smarter ways of doing things 👍 didn't know np.round took that parameter!

andrew-d-king · 2026-05-15T13:10:12Z

+
+        return sampler_call
+
+    def _format_parameter_list(


Maybe not for this PR, but ultimately I think that a parameter class should be written, with a format parameter (basically sig figs, for float parameters). It's a bit ugly to do it ad hoc, as I've written it.

andrew-d-king · 2026-05-15T13:12:08Z

+        filename = title
+        for bad_symbol in "/: ;,":
+            filename = filename.replace(bad_symbol, "_")
+        fig.savefig(Path(os.getcwd()) / 'figures' / f"{filename}.png")


All the figures should be tight_layout(), then saved, then shown. It's inconsistent right now.

andrew-d-king · 2026-05-15T13:14:05Z

+plt.show()
+
+# Now we will analyze the kink-kink correlator for the fastest anneals (5ns)
+fig3, ax3 = plt.subplots(1, 2, figsize=(10, 8))


It should be noted that running only five iterations will give poor results for ckk, which is fairly sensitive and requires a converged shim. It also requires more data. But we don't want to put too much QPU usage in a mere example, so I think it is best to just point this out clearly.

andrew-d-king · 2026-05-15T13:22:53Z

+
+        ax = axes[2, 0]
+        ax.plot(cshim[0])
+        ax.set_title(f"Coupler shim, t_a={ANNEAL_TIMES[0]:.3f}μs")


Suggested change

ax.set_title(f"Coupler shim, t_a={ANNEAL_TIMES[0]:.3f}μs")

ax.set_title(f"Coupler shim, $t_a=${ANNEAL_TIMES[0]:.3f}μs")

andrew-d-king · 2026-05-15T13:23:07Z

+
+        ax = axes[2, 1]
+        ax.plot(cshim[1])
+        ax.set_title(f"Coupler shim, t_a={ANNEAL_TIMES[1]:.3f}μs")


Suggested change

ax.set_title(f"Coupler shim, t_a={ANNEAL_TIMES[1]:.3f}μs")

ax.set_title(f"Coupler shim, $t_a=${ANNEAL_TIMES[1]:.3f}μs")

andrew-d-king · 2026-05-15T13:23:21Z

+
+        ax = axes[2, 2]
+        ax.plot(cshim[6])
+        ax.set_title(f"Coupler shim, t_a={ANNEAL_TIMES[-1]:.3f}μs")


Suggested change

ax.set_title(f"Coupler shim, t_a={ANNEAL_TIMES[-1]:.3f}μs")

ax.set_title(f"Coupler shim, $t_a=${ANNEAL_TIMES[-1]:.3f}μs")

andrew-d-king · 2026-05-15T14:48:50Z

+        config = experiment.FastAnnealExperimentConfig(
+            energy_scale=energy_scale,
+            coupler_shim_step=0.05,
+            flux_bias_shim_step=1e-6,


Suggested change

flux_bias_shim_step=1e-6,

flux_bias_shim_step=3e-6,

andrew-d-king · 2026-05-15T14:49:37Z

+        # Make parameter list. We will only vary anneal time.
+        parameter_list = [{"anneal_time": time} for time in ANNEAL_TIMES]
+
+        for _ in range(20): #TODO clean this up?


Suggested change

for _ in range(20): #TODO clean this up?

while True:

SebastianGitt added 2 commits March 13, 2026 13:42

Add lattice-utils to dwave-experimental

7dac34b

Add unit tests

23938bf

-move initialization of num_spins to subclass

thisac reviewed Apr 24, 2026

View reviewed changes

SebastianGitt added 7 commits May 7, 2026 11:08

Improve docstrings and formatting

759eca5

-added more comprehensive docstrings -fixed formatting issues

Improve docstrings and formatting

b6bf8f6

-added more comprehensive docstrings -fixed formatting issues

Merge branch 'add-lattice-utils' of https://github.com/SebastianGitt/…

a5ae37d

…dwave-experimental into add-lattice-utils

Merge remote-tracking branch 'sebastians-fork/main' into add-lattice-…

0aa1bff

…utils

Remove temporary automorphism file and additional cleanup

8122131

-removed spin reversal transform functionality -uses automorphism module already in dwave-experimental -uses data classes for experiment configs

Remove temporary automorphism file and additional cleanup

0ef6a52

-removed spin reversal transform functionality -uses automorphism module already in dwave-experimental -uses data classes for experiment configs -formatted using black

Format using black

ab9348a

SebastianGitt marked this pull request as ready for review May 14, 2026 17:50

andrew-d-king suggested changes May 15, 2026

View reviewed changes



		def confidence_interval(array: NDArray, width: float = 0.95) -> tuple[float, float, float]:
		"""Ravel and take the quantiles; return median and error bar lengths."""

	seed: int \| None = None,
	rng: np.random.Generator \| None = None,

		if not hasattr(self, "num_spins"):
		raise AttributeError(f"{type(self).__name__} subclass must initialize self.num_spins")

	ax.set_title(f"Coupler shim, t_a={ANNEAL_TIMES[0]:.3f}μs")
	ax.set_title(f"Coupler shim, $t_a=${ANNEAL_TIMES[0]:.3f}μs")

Conversation

SebastianGitt commented Mar 13, 2026

Uh oh!

thisac left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

andrew-d-king left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants