DAGCircuit: Add get_causal_cone method #10325

raynelfss · 2023-06-23T04:37:48Z

Summary

The causal node of a qubit is the set of all the qubits that affect the resulting output of the aforementioned qubit. This PR aims to add the get_causal_node method to DAGCircuit to obtain the causal node of a qubit in a circuit.

A couple of helper methods have also been added:

get_qubit_input_node: to get the input node of a certain qubit in the circuit.
get_qubit_output_node: to get the output node of a certain qubit in the circuit.

Details and comments

The implementation of these methods may not be the most optimal one, so it is open to suggestions from the community.

- Also added `get_qubit_input_node` and `get_qubit_output_node`.

qiskit-bot · 2023-06-23T04:37:51Z

Thank you for opening a new pull request.

Before your PR can be merged it will first need to pass continuous integration tests and be reviewed. Sometimes the review process can be slow, so please be patient.

While you're waiting, please feel free to review other open PRs. While only a subset of people are authorized to approve pull requests for merging, everyone is encouraged to review open pull requests. Doing reviews helps reduce the burden on the core team and helps make the project's code better for everyone.

One or more of the the following people are requested to review this:

@Qiskit/terra-core

- Type checking was causing strange behavior during linting.

coveralls · 2023-06-23T09:50:12Z

Pull Request Test Coverage Report for Build 5562839866

16 of 17 (94.12%) changed or added relevant lines in 1 file are covered.
20 unchanged lines in 4 files lost coverage.
Overall coverage decreased (-0.01%) to 86.053%

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
qiskit/dagcircuit/dagcircuit.py	16	17	94.12%

Files with Coverage Reduction	New Missed Lines	%
qiskit/transpiler/passes/synthesis/unitary_synthesis.py	1	90.33%
qiskit/extensions/quantum_initializer/squ.py	2	80.0%
crates/qasm2/src/lex.rs	5	91.14%
crates/qasm2/src/parse.rs	12	97.11%

Totals
Change from base Build 5559766630:	-0.01%
Covered Lines:	72273
Relevant Lines:	83987

💛 - Coveralls

- Replace lists with deque for the iteration.

CLAassistant · 2023-06-26T20:23:55Z

All committers have signed the CLA.

CLAassistant · 2023-06-26T20:23:55Z

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
2 out of 3 committers have signed the CLA.

✅ raynelfss
✅ danielleodigie
❌ Raynel Sanchez

Raynel Sanchez seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

danielleodigie

I think this looks good to me!

mtreinish

This is a great start, thanks for getting it going. I left a few inline comments on some small cleanups you can make to the interface to make the implementation a bit simpler and also more consistent with the rest of the DAGCircuit API.

qiskit/dagcircuit/dagcircuit.py

mtreinish

Thanks for the fast updates, just a couple of small comments inline, mostly about doc formatting.

qiskit/dagcircuit/dagcircuit.py

releasenotes/notes/add-dag-causal-cone-5a19311e40fbb3af.yaml

qiskit/dagcircuit/dagcircuit.py

mtreinish

I think this is basically ready now, just a couple of small inline comments, but after that I think it's ready to merge

mtreinish · 2023-07-17T16:30:15Z

qiskit/dagcircuit/dagcircuit.py

+        Returns:
+            Set[Qubit]: The set of qubits whose interactions affect ``qubit``.
+        """
+        # Retrieve the output node and the qubit


Suggested change

# Retrieve the output node and the qubit

# Retrieve the output node from the qubit

mtreinish · 2023-07-17T16:31:13Z

qiskit/dagcircuit/dagcircuit.py

+                # Check if there are any qubits in common and that the operation is not a barrier.
+                if (
+                    len(qubit_set.intersection(qubits_to_check)) > 0
+                    and not node_to_check.op.name == "barrier"


Suggested change

and not node_to_check.op.name == "barrier"

and node_to_check.op.name != "barrier"

mtreinish · 2023-07-17T16:32:07Z

qiskit/dagcircuit/dagcircuit.py

+                if (
+                    len(qubit_set.intersection(qubits_to_check)) > 0
+                    and not node_to_check.op.name == "barrier"
+                ):


We probably want to add a check on: not getattr(node_to_check.op, "_directive") which is an expansion of the barrier check so it'll exclude other directives too, like snapshot from aer, etc.

jakelishman

I think your code will already handle this correctly, but it'd be good to add a test that this function will include qubits in the causal cone that interact with the target qubit only via a classical condition.

For example,

qc = QuantumCircuit(2, 1)
qc.x(0)
qc.measure(0, 0)
qc.x(1).c_if(0, True)

In that circuit, qubit 0 is in the causal cone of qubit 1, but it never appears in a 2q quantum operation. I'm fairly sure your code already handles this right, but it'd be good to have the extra test.

raynelfss · 2023-07-17T17:04:47Z

I think your code will already handle this correctly, but it'd be good to add a test that this function will include qubits in the causal cone that interact with the target qubit only via a classical condition.

Well, I just tried this with the code and it doesn't seem to handle it well. Since the controlled operation never happens directly in a multiqubit operation and the condition is happening based on a classical condition, it will ignore it as the function only considers qubits and not classical bits.

This is an interesting test case though, and raises a great question of how direct do interactions have to be for it to be considered as part of the causal cone. In this case, since the output of qubit 0 affects the output of qubit 1, it should be part of its causal cone.

I will try to work that into the code, thank you for bringing it up @jakelishman.

jakelishman · 2023-07-17T17:10:42Z

Thinking about it, it actually gets very hairy very quickly, because with clbits there's a distinction between "writes" and "reads", which the DAGCircuit doesn't represent right now. For example, if I change my example to:

qc = QuantumCircuit(3, 1)  # Note the extra qubit
qc.x(0)
qc.measure(0, 0)
qc.x(1).c_if(0, True)
qc.x(2).c_if(0, True)

Now the qc.x(1) operation will be a DAGCircuit predecessor of the x(2) operation, but qubit 1 isn't in the causal cone of qubit 2, because it doesn't write to the wire that orders them.

I'm not sure that it's going to be reasonable to handle this at the moment; the DAGCircuit just doesn't give us the information we need right now to determine it. Matt might have an opinion on what the best API choice for this is.

mtreinish · 2023-07-17T18:11:08Z

TBH, I'm thinking it might be better to call this quantum_causal_cone (I feel like this is sounding more and more scifi) and make this explicitly for quantum data dependency only and document it as such. At least then we're punting the question of classical data dependency for now, because at the least the primary use cases for this are only about the quantum side. The question for me then is whether we do the normal dagcircuit thing of treating a .condition as breaking the data dependency (i.e. collect_runs() which terminates the run on a controlled gate) explicitly. Or just documenting it as not being accounted for in the output and not trying to handle it explicitly. In general I think just making explicitly only for quantum data dependencies and documenting it clearly will be enough.

mtreinish · 2023-07-18T18:18:45Z

qiskit/dagcircuit/dagcircuit.py

+        the causal cone of a qubit can be useful when debugging faulty circuits, as it can
+        help identify which wire(s) may be causing the problem.
+
+        This method does not consider classical to quantum influences through conditionals or


I'm finding this a bit hard to follow, maybe reword this a bit to something like?:

Suggested change

This method does not consider classical to quantum influences through conditionals or

This method does not consider any classical data dependency in the ``DAGCircuit``, classical

bit wires are ignored for the purposes of building the causal cone.

mtreinish · 2023-07-18T18:20:01Z

qiskit/dagcircuit/dagcircuit.py

+                    qubits_to_check = qubits_to_check.union(qubit_set)
+            # For each predecessor of the current node, filter input/output nodes,
+            # also make sure it has at least one qubit in common. Then append.
+            for node in self.predecessors(node_to_check):


If we want to hold to the statement of ignoring the classical wires should we change this to?:

Suggested change

for node in self.predecessors(node_to_check):

for node in self.quantum_predecessors(node_to_check):

(and above as well).

mtreinish

LGTM now, thanks for all the quick updates.

mtreinish · 2023-07-18T19:38:19Z

qiskit/dagcircuit/dagcircuit.py

+        # Add the qubit to the causal cone.
+        qubits_to_check = {qubit}
+        # Add predecessors of output node to the queue.
+        queue = deque(self.predecessors(output_node))


This could be quantum predecessors too, but if the input is a Qubit then there is no way that output_node could have a classical wire into it.

raynelfss added 3 commits June 23, 2023 00:28

Feat: Add get_causal_node to DAGCircuit:

1960f34

- Also added `get_qubit_input_node` and `get_qubit_output_node`.

Test: Added tests to dagcircuit.py

7f633e7

Docs: Added release note

9f98755

raynelfss requested a review from a team as a code owner June 23, 2023 04:37

qiskit-bot added the Community PR PRs from contributors that are not 'members' of the Qiskit repo label Jun 23, 2023

Chore: Remove type-checking in dagcircuit.py

5df78a4

- Type checking was causing strange behavior during linting.

mtreinish added Changelog: New Feature Include in the "Added" section of the changelog Intern PR PR submitted by IBM Quantum interns labels Jun 23, 2023

raynelfss and others added 3 commits June 23, 2023 10:10

Merge branch 'main' into dagcircuit/causal-cone

e7357ed

Added changes to speed up get_causal_cone (#1)

b599f84

- Replace lists with deque for the iteration.

Merge branch 'main' into dagcircuit/causal-cone

7afeaca

raynelfss force-pushed the dagcircuit/causal-cone branch from 826da16 to 7afeaca Compare June 26, 2023 20:33

raynelfss added 4 commits June 26, 2023 16:34

Docs: Modify docstring and release note

23d4bf4

Merge branch 'Qiskit:main' into dagcircuit/causal-cone

9b67119

Merge branch 'main' into dagcircuit/causal-cone

be9df27

Merge branch 'main' into dagcircuit/causal-cone

6d57296

jakelishman removed the Community PR PRs from contributors that are not 'members' of the Qiskit repo label Jun 30, 2023

raynelfss added 2 commits July 7, 2023 12:46

Merge branch 'main' into dagcircuit/causal-cone

e201796

Fix: Wrong comparison in _get_input_output_node

7959ffc

mtreinish added this to the 0.25.0 milestone Jul 10, 2023

danielleodigie reviewed Jul 11, 2023

View reviewed changes

mtreinish reviewed Jul 13, 2023

View reviewed changes

raynelfss added 4 commits July 13, 2023 21:54

Merge branch 'main' into dagcircuit/causal-cone

2bcced6

Remove: input and output node methods.

373673c

Lint: Fixed formatting

7cdb264

Docs: Fixed release-note

4e80937

mtreinish reviewed Jul 14, 2023

View reviewed changes

raynelfss added 3 commits July 15, 2023 10:53

Docs: Fixed docstring and release note.

59fb330

Fix: Output map double-lookup.

16350c5

Merge branch 'main' into dagcircuit/causal-cone

9158c03

mtreinish reviewed Jul 17, 2023

View reviewed changes

jakelishman reviewed Jul 17, 2023

View reviewed changes

raynelfss added 4 commits July 17, 2023 15:51

Docs: Fix inline comments.

d10eb8e

Test: Added test for circuits with barriers

d129ffc

Refactor: rename to quantum_causal_cone

b924ea2

Merge branch 'main' into dagcircuit/causal-cone

d66e2f0

mtreinish reviewed Jul 18, 2023

View reviewed changes

raynelfss added 2 commits July 18, 2023 14:44

FIx: Use quantum_sucessors and docstring

65109c3

Merge branch 'main' into dagcircuit/causal-cone

005f077

mtreinish approved these changes Jul 18, 2023

View reviewed changes

mtreinish enabled auto-merge July 18, 2023 19:39

mtreinish added this pull request to the merge queue Jul 18, 2023

Merged via the queue into Qiskit:main with commit 90c2fe6 Jul 18, 2023

raynelfss deleted the dagcircuit/causal-cone branch July 21, 2023 02:13

alexanderivrii mentioned this pull request Jun 26, 2024

improving quantum_causal_cone method in python #12668

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

DAGCircuit: Add get_causal_cone method #10325

DAGCircuit: Add get_causal_cone method #10325

raynelfss commented Jun 23, 2023

qiskit-bot commented Jun 23, 2023

coveralls commented Jun 23, 2023 •

edited

Loading

CLAassistant commented Jun 26, 2023 •

edited

Loading

CLAassistant commented Jun 26, 2023

danielleodigie left a comment

mtreinish left a comment

mtreinish left a comment

mtreinish left a comment

mtreinish Jul 17, 2023

mtreinish Jul 17, 2023

mtreinish Jul 17, 2023

jakelishman left a comment •

edited

Loading

raynelfss commented Jul 17, 2023

jakelishman commented Jul 17, 2023

mtreinish commented Jul 17, 2023

mtreinish Jul 18, 2023

mtreinish Jul 18, 2023

mtreinish left a comment

mtreinish Jul 18, 2023

	# Retrieve the output node and the qubit
	# Retrieve the output node from the qubit

	and not node_to_check.op.name == "barrier"
	and node_to_check.op.name != "barrier"

	This method does not consider classical to quantum influences through conditionals or
	This method does not consider any classical data dependency in the ``DAGCircuit``, classical
	bit wires are ignored for the purposes of building the causal cone.

	for node in self.predecessors(node_to_check):
	for node in self.quantum_predecessors(node_to_check):

DAGCircuit: Add get_causal_cone method #10325

DAGCircuit: Add get_causal_cone method #10325

Conversation

raynelfss commented Jun 23, 2023

Summary

Details and comments

qiskit-bot commented Jun 23, 2023

coveralls commented Jun 23, 2023 • edited Loading

Pull Request Test Coverage Report for Build 5562839866

💛 - Coveralls

CLAassistant commented Jun 26, 2023 • edited Loading

CLAassistant commented Jun 26, 2023

danielleodigie left a comment

Choose a reason for hiding this comment

mtreinish left a comment

Choose a reason for hiding this comment

mtreinish left a comment

Choose a reason for hiding this comment

mtreinish left a comment

Choose a reason for hiding this comment

mtreinish Jul 17, 2023

Choose a reason for hiding this comment

mtreinish Jul 17, 2023

Choose a reason for hiding this comment

mtreinish Jul 17, 2023

Choose a reason for hiding this comment

jakelishman left a comment • edited Loading

Choose a reason for hiding this comment

raynelfss commented Jul 17, 2023

jakelishman commented Jul 17, 2023

mtreinish commented Jul 17, 2023

mtreinish Jul 18, 2023

Choose a reason for hiding this comment

mtreinish Jul 18, 2023

Choose a reason for hiding this comment

mtreinish left a comment

Choose a reason for hiding this comment

mtreinish Jul 18, 2023

Choose a reason for hiding this comment

coveralls commented Jun 23, 2023 •

edited

Loading

CLAassistant commented Jun 26, 2023 •

edited

Loading

jakelishman left a comment •

edited

Loading