feat: Add CopyableExpressionAST

[lmondada]

I want to define a Subcircuit as a combination of Intervals on resource paths (the parts with linear, non-copyable data), and ASTs of fully classical, copyable data.

• PR feat!: Interval is independent of resource IDs and scope position #1205 was opened to refine the notion of Interval
• this PR creates the CopyableExpressionAST type that corresponds to ASTs of classical data in hugr graphs.

[codecov]

Codecov Report

❌ Patch coverage is 93.53234% with 13 lines in your changes missing coverage. Please review.
✅ Project coverage is 78.97%. Comparing base (d34b013) to head (c7278d6).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
tket/src/subcircuit/expression.rs	94.35%	9 Missing and 1 partial ⚠️
tket/src/resource/scope.rs	87.50%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1209      +/-   ##
==========================================
+ Coverage   78.77%   78.97%   +0.20%     
==========================================
  Files         153      154       +1     
  Lines       18910    19111     +201     
  Branches    17808    18009     +201     
==========================================
+ Hits        14896    15093     +197     
- Misses       3070     3076       +6     
+ Partials      944      942       -2     

Flag	Coverage Δ
python	92.65% <ø> (ø)
qis-compiler	68.40% <ø> (ø)
rust	78.59% <93.53%> (+0.21%)	⬆️

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.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[acl-cqc]

I've suggested a lot of renamings and minor stuff, but nothing major, so I guess I'm happy for this to go in, but with the proviso...

...that I am still not certain whether we shouldn't have a single SiblingSubgraph for all the classical computation associated with a set of intervals rather than many potentially-overlapping ones...maybe I would need to see some examples

[lmondada]

not certain whether we shouldn't have a single SiblingSubgraph for all the classical computation associated with a set of intervals rather than many potentially-overlapping ones...

I agree that the fact that multiple copyable expressions might have overlapping nodes is not great. That being said

1. your suggestion would require removing the convexity assumption (and enforcement) of [SiblingSubgraph] in hugr. This move would be quite breaking and should be thought through well. Might be something to add to clean-up november+december 😅 I agree that if non-convex SiblingSubgraphs existed, we could design this differently
2. You'll see in the upcoming Subcircuit PR that these (overlapping) copyable expressions are not stored within the subcircuit -- they are constructed on the fly from the set of inputs of the subcircuit. In other words, they are not part of the data of the subcircuit, but are much more like computable properties of the subcircuit. Thus, the fact that overlapping expressions can be produced depending on which copyable values we are interested in is not so concerning.

Overall, I agree that there are further follow-ups and clean-ups that should be considered and executed, but this is a step in the right direction: it's helping construct a layer of abstraction above SiblingSubgraph, which is badly needed.

[lmondada]

Thanks a lot Alan for the review!! Will now go through your comments and merge this in.