Refactor OpenQASM 2 exporter #9953
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This rewrites the handling of instructions in the OpenQASM 2 exporter in several places. First, the definition and the call site are handled at the same time. Previously, these were handled at separate points and required two name munging steps, which mostly just needed to hope that they both converged on the same name. Instead, one singular point handles both, so there are no issues with name mismatches. In a similar vein, this now ensures that the two places will not disagree about when a definition has already been seen before; previously, duplicate definitions could be "used" but not defined. Second, the gate-definition pass is now made properly recursive, so the exact same code is responsible for defining gates that appear in the body of the main circuit, and of custom definitions, including sharing the duplication detection. Previously, two custom definitions that both used the same gate could cause duplicate definitions to appear with incorrect names. Unifying this handling also has the natural consequence of correctly ordering _all_ gate definitions, whereas before some inner gates could be defined after their first use. Inside the call-site/definition-site handling, this refactor separates those into two separate objects. This massively simplified: - handling known-good Qiskit standard gates to have special-cased support for always exporting properly parametrised definitions; - handling `UnitaryGate` and `Permutation` specially, so they no longer need to output special OQ2 strings, but can just define a custom object to use as the source for both call-site and definition site; - handling gates with no definitions as `opaque`.
jakelishman
added
Changelog: Bugfix
|
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: |
Pull Request Test Coverage Report for Build 4721104772
💛 - Coveralls |
1ucian0
reviewed
Apr 15, 2023
| operation.qasm(), | ||
| ",".join([bit_labels[j] for j in instruction.qubits + instruction.clbits]), | ||
| bits_qasm = ",".join( | ||
| bit_labels[j] for j in itertools.chain(instruction.qubits, instruction.clbits) |
There was a problem hiding this comment.
why not instruction.qubits + instruction.clbits as before?
There was a problem hiding this comment.
Oh, no particular reason - I probably just didn't notice I'd done it during the rewrite. itertools.chain is the normal idiom I personally use, so I just wouldn't have thought about it.
1ucian0
reviewed
Apr 15, 2023
1ucian0
previously approved these changes
Apr 15, 2023
1ucian0
reviewed
Apr 15, 2023
Co-authored-by: Luciano Bello <bel@zurich.ibm.com>
1ucian0
approved these changes
Apr 17, 2023
This was referenced May 4, 2023
king-p3nguin
pushed a commit
to king-p3nguin/qiskit-terra
that referenced
this pull request
May 22, 2023
* Refactor OpenQASM 2 exporter This rewrites the handling of instructions in the OpenQASM 2 exporter in several places. First, the definition and the call site are handled at the same time. Previously, these were handled at separate points and required two name munging steps, which mostly just needed to hope that they both converged on the same name. Instead, one singular point handles both, so there are no issues with name mismatches. In a similar vein, this now ensures that the two places will not disagree about when a definition has already been seen before; previously, duplicate definitions could be "used" but not defined. Second, the gate-definition pass is now made properly recursive, so the exact same code is responsible for defining gates that appear in the body of the main circuit, and of custom definitions, including sharing the duplication detection. Previously, two custom definitions that both used the same gate could cause duplicate definitions to appear with incorrect names. Unifying this handling also has the natural consequence of correctly ordering _all_ gate definitions, whereas before some inner gates could be defined after their first use. Inside the call-site/definition-site handling, this refactor separates those into two separate objects. This massively simplified: - handling known-good Qiskit standard gates to have special-cased support for always exporting properly parametrised definitions; - handling `UnitaryGate` and `Permutation` specially, so they no longer need to output special OQ2 strings, but can just define a custom object to use as the source for both call-site and definition site; - handling gates with no definitions as `opaque`. * Style tweak Co-authored-by: Luciano Bello <bel@zurich.ibm.com> --------- Co-authored-by: Luciano Bello <bel@zurich.ibm.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This rewrites the handling of instructions in the OpenQASM 2 exporter in several places.
First, the definition and the call site are handled at the same time. Previously, these were handled at separate points and required two name munging steps, which mostly just needed to hope that they both converged on the same name. Instead, one singular point handles both, so there are no issues with name mismatches. In a similar vein, this now ensures that the two places will not disagree about when a definition has already been seen before; previously, duplicate definitions could be "used" but not defined.
Second, the gate-definition pass is now made properly recursive, so the exact same code is responsible for defining gates that appear in the body of the main circuit, and of custom definitions, including sharing the duplication detection. Previously, two custom definitions that both used the same gate could cause duplicate definitions to appear with incorrect names. Unifying this handling also has the natural consequence of correctly ordering all gate definitions, whereas before some inner gates could be defined after their first use.
Inside the call-site/definition-site handling, this refactor separates those into two separate objects. This massively simplified:
UnitaryGateandPermutationspecially, so they no longer need to output special OQ2 strings, but can just define a custom object to use as the source for both call-site and definition site;opaque.Details and comments
This fixes a whole slew of long-standing bugs, albeit not perfectly. The general case of user-defined parametrised gates are still not handled well, but Qiskit's standard gates now generally are. Even in the worst cases, the export should now be much safer and produce rather less invalid OQ2 code than it did before.
To do with duplicate definitions:
RYYGatein subcircuit (with inverse)QasmError: "Duplicate declaration for gate 'ryy'#7771QuantumCircuit.qasmcannot handle nested gates with different parameters #9805To do with the old
UnitaryGate.qasm:CSwapGateandDCXGate:Cannot find gate definition for 'unitary...'#7772CSwapGateandCCXGate:Cannot find gate definition for 'unitary...'#8222To do with bad orderings of definitions:
Cannot find gate definition for 'rcccx'#7773A bit more of a mix:
opaque#5036