Dont batch unless separate samplers are used for each subcircuit (#333)

* Don't batch unnecessarily

* mypy

* black

* Missing varname

* fix bug

* Clean up code

* Improve comments

* release note

* release note

* Update batch-by-sampler-c4ae836df9997b1d.yaml

* Update batch-by-sampler-c4ae836df9997b1d.yaml

* Update batch-by-sampler-c4ae836df9997b1d.yaml

* Bump Python version in Dockerfile to 3.11 (#331)

Now that CKT supports Python 3.11, we might as well use the latest version in the Dockerfile.

* Fix formatting of example in `reduce_bitstrings` docstring (#332)

* Implement cutting of general 2-qubit unitaries (#302)

* Add support for `SwapGate`

* Reorder terms

[ci skip]

* Add missing terms

* DRY the coefficients

* Fix coverage

* Add support for `iSwapGate`

* Fix black

* Add to release note

* Fix type hint

* Gates without parameters are nicer to work with

and can be singletons, one day!

* Remove a line

* Add comments describing channels

* `_copy_unique_sublists`

* Add `DCXGate`

* Tweak

* Implement cutting of general 2-qubit unitaries

Builds on #294.  Closes #186.

* Add tests of additional gates

* Fix type annotation

* Add explanatory comments

* `supported_gates()` -> `explicitly_supported_gates()`

* Add to references

* Improved error message and test about `to_matrix` conversion failing

* Add xref to `QPDBasis` in docstrings

* Add `qpdbasis_from_gate` to Sphinx build

* Make `explicitly_supported_gates` private and remove its release note

It's not clear that this function remains useful now that we
support essentially all 2-qubit gates.  If we find a use for it
in the future, we can re-introduce it (or something like it) as a
public interface.

* Fix intersphinx link

* Release note

* Update qpd.py: remove extraneous `from None`

* Improve the instructions regarding pandoc (#336)

I've also tried to make the developer documentation easier to discover

* Make the repository link more obvious from the Sphinx build (#338)

* Make the repository link more obvious from the Sphinx build

* Enable "edit" link in the header

* Add comment

* Add README badge linking to stable documentation (#339)

* Add docs badge to link to stable docs

* Add ruff badge

* Remove ruff

* Update README.md (#340)

Making directions on opening the docs a little more succinct.

* Change var name

* Add SECURITY.md (#337)

* Implement wire cutting as a two-qubit instruction (#174)

* Implement wire cutting as a two-qubit instruction

* Update type annotation

* s/gate/instruction/

* Add overhead test for `Move` instruction

* Add wire cutting tutorial

* Add `Move` to Sphinx build

* Doc updates suggested by Caleb

* Add release note and link to new tutorial

* Clarify wording

following #174 (comment)

* Improvements to `Move` docstring

* Use svg as the plot format

This avoids pixelation on high-dpi displays

* Remove unnecessary uses of `CircuitInstruction`

https://github.com/Qiskit-Extensions/circuit-knitting-toolbox/pull/174/files#r1278067109

* The notebook tests should ignore any files that crop up in `docs/_build`

`matplotlib.sphinxext.plot_directive` likes to leave python files there

---------

Co-authored-by: Jim Garrison <garrison@ibm.com>
Co-authored-by: Ibrahim Shehzad <75153717+IbrahimShehzad@users.noreply.github.com>

circuit_knitting/cutting/cutting_evaluation.py

@@ -128,30 +128,47 @@ def execute_experiments(
         num_samples,
     )
 
-    # Create a list of samplers -- one for each qubit partition
-    num_partitions = len(subexperiments[0])
+    # Create a list of samplers to use -- one for each batch
     if isinstance(samplers, BaseSampler):
-        samplers_by_partition = [samplers] * num_partitions
+        samplers_by_batch = [samplers]
+        batches = [
+            [
+                sample[i]
+                for sample in subexperiments
+                for i in range(len(subexperiments[0]))
+            ]
+        ]
     else:
-        samplers_by_partition = [samplers[key] for key in sorted(samplers.keys())]
+        samplers_by_batch = [samplers[key] for key in sorted(samplers.keys())]
+        batches = [
+            [sample[i] for sample in subexperiments]
+            for i in range(len(subexperiments[0]))
+        ]
 
-    # Run each partition's sub-experiments
-    quasi_dists_by_partition = [
+    # There should be one batch per input sampler
+    assert len(samplers_by_batch) == len(batches)
+
+    # Run each batch of sub-experiments
+    quasi_dists_by_batch = [
         _run_experiments_batch(
-            [sample[i] for sample in subexperiments],
-            samplers_by_partition[i],
+            batches[i],
+            samplers_by_batch[i],
         )
-        for i in range(num_partitions)
+        for i in range(len(samplers_by_batch))
     ]
 
-    # Reformat the counts to match the shape of the input before returning
-    num_unique_samples = len(subexperiments)
+    # Build the output data structure to match the shape of input subexperiments
     quasi_dists: list[list[list[tuple[dict[str, int], int]]]] = [
-        [] for _ in range(num_unique_samples)
+        [] for _ in range(len(subexperiments))
     ]
-    for i in range(num_unique_samples):
-        for partition in quasi_dists_by_partition:
-            quasi_dists[i].append(partition[i])
+    count = 0
+    for i in range(len(subexperiments)):
+        for j in range(len(subexperiments[0])):
+            if len(samplers_by_batch) == 1:
+                quasi_dists[i].append(quasi_dists_by_batch[0][count])
+                count += 1
+            else:
+                quasi_dists[i].append(quasi_dists_by_batch[j][i])
 
     return CuttingExperimentResults(quasi_dists, coefficients)
 
@@ -327,18 +344,14 @@ def _run_experiments_batch(
     quasi_dists_flat = sampler.run(experiments_flat).result().quasi_dists
 
     # Reshape the output data to match the input
-    if len(subexperiments) == 1:
-        quasi_dists_reshaped = np.array([quasi_dists_flat])
-        num_qpd_bits = np.array([num_qpd_bits_flat])
-    else:
-        # We manually build the shape tuple in second arg because it behaves strangely
-        # with QuantumCircuits in some versions. (e.g. passes local pytest but fails in tox env)
-        quasi_dists_reshaped = np.reshape(
-            quasi_dists_flat, (len(subexperiments), len(subexperiments[0]))
-        )
-        num_qpd_bits = np.reshape(
-            num_qpd_bits_flat, (len(subexperiments), len(subexperiments[0]))
-        )
+    quasi_dists_reshaped: list[list[QuasiDistribution]] = [[] for _ in subexperiments]
+    num_qpd_bits: list[list[int]] = [[] for _ in subexperiments]
+    count = 0
+    for i, subcirc in enumerate(subexperiments):
+        for j in range(len(subcirc)):
+            quasi_dists_reshaped[i].append(quasi_dists_flat[count])
+            num_qpd_bits[i].append(num_qpd_bits_flat[count])
+            count += 1
 
     # Create the counts tuples, which include the number of QPD measurement bits
     quasi_dists: list[list[tuple[dict[str, float], int]]] = [

releasenotes/notes/batch-by-sampler-c4ae836df9997b1d.yaml

@@ -0,0 +1,5 @@
+---
+upgrade:
+  - |
+    :func:`~circuit_knitting.cutting.execute_experiments` no longer creates separate jobs for each subcircuit by default.
+    Now, separate jobs are only created if separate :class:`~qiskit.primitives.BaseSampler` instances are provided for each circuit partition.