Improving docstrings in final qubit mapping, w/ {logical: physical}

[GuillermoAbadLopez]

PR addressing Issue: #1562

• Changes incorrect CircuitMap.final_layout() docstring from physical-logical to logical-physical
• Changes all logical-physical by {logical: physical}, for showing more explicitly which are the keys/values

Extra for reviewers:

• Please check the first change from above is correct, since I didn't went too deeply to check it. Instead I assumed it has to be that way based on the contradiction there was, and checking the next method names it calls 👍

Checklist:

• Reviewers confirm new code works as expected.
• Tests are passing.
• Coverage does not decrease.
• Documentation is updated.

[alecandido]

Nothing to amend, we can merge.

---

General comment, beyond this PR

Just notice that the logical and physical adjectives were also a poor choice, since most often they are just intermediate steps, so it's just a way of handling a permutation.

In principle, we could return just a permutation as a sequence of permuted indices.
However, the common output of each step is a transformed circuit, which may be remodeled including a permutation, or boil down to just that. But it may even contain no permutation at all, and just operate on the gates (as in the case of the Unroller).

So, I would discourage any further focus on permutations. Including the current final_layout, which is only sometimes present, and leading to patterns like the following

qibo/src/qibo/transpiler/pipeline.py

Lines 84 to 101 in d32efb8

	for transpiler_pass in self.passes:
	if isinstance(transpiler_pass, Optimizer):
	transpiler_pass.connectivity = self.connectivity
	circuit = transpiler_pass(circuit)
	elif isinstance(transpiler_pass, Placer):
	transpiler_pass.connectivity = self.connectivity
	final_layout = transpiler_pass(circuit)
	elif isinstance(transpiler_pass, Router):
	transpiler_pass.connectivity = self.connectivity
	circuit, final_layout = transpiler_pass(circuit)
	elif isinstance(transpiler_pass, Unroller):
	circuit = transpiler_pass(circuit)
	else:
	raise_error(
	TranspilerPipelineError,
	f"Unrecognised transpiler pass: {transpiler_pass}",
	)
	return circuit, final_layout

Just relying the Circuit.wire_names will embed the optional permutation in the resulting Circuit.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.68%. Comparing base (d32efb8) to head (6145cc9).
Report is 2 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1563   +/-   ##
=======================================
  Coverage   99.68%   99.68%           
=======================================
  Files          76       76           
  Lines       11280    11280           
=======================================
  Hits        11244    11244           
  Misses         36       36           

Flag	Coverage Δ
unittests	99.68% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[GuillermoAbadLopez]

PR itself:

Nothing to amend, we can merge.

Cool to hear 🙌

---

Beyond the PR:

Just notice that the logical and physical adjectives were also a poor choice, since most often they are just intermediate steps, so it's just a way of handling a permutation.

Yes, and qiskit has changed all these names several times not too long ago, so I would say the industry hasn't yet settled on anything, and it can be tricky to keep up with what each permutation represents for each different SDK.. 😅

In principle, we could return just a permutation as a sequence of permuted indices.

Precisely!

So, I would discourage any further focus on permutations. Including the current final_layout, which is only sometimes present, and leading to patterns like the following
[source_code]
Just relying the Circuit.wire_names will embed the optional permutation in the resulting Circuit.

That makes sense, I'll try it, thanks!