refactor: add HugrView::scheduling_graph, deprecate region_portgraph by acl-cqc · Pull Request #3015 · Quantinuum/hugr

acl-cqc · 2026-04-02T11:53:36Z

This paves the way for #2946, but does not solve. For that, we will need to add "fake" edges: for each nonlocal edge, from the source to the sibling that is ancestor of the target. (Currently stored in Hugr as order edges but these will be removed). However, it will not be possible to add these fake edges to the region_portgraph: being a portgraph (view of hugr), there is no port for the fake edges.

Instead this PR introduces a new method scheduling_graph returning a new SchedulingGraph structure: a generic petgraph, which augments a view of the Hugr with an explicit list of extra edges. In this PR, said list is empty, as the Hugr still contains the order edges necessary to ensure topological orderings are sound. #2951 removes the order edges from Hugr and generates a list of those required as part of the SchedulingGraph.

This PR includes moving sibling_subgraph.rs tests into a separate tests.rs, and also copies the TopoConvexChecker from portgraph (reparametrized to work on a petgraph).

Add some tests demonstrating existing behaviour of SiblingSubgraph and scheduling_graph with order edges, these are then updated in #2951 to show the (i.e. breaking) changes in behaviour.

…Topo

…woot!)

…Graph

codecov · 2026-04-05T09:42:40Z

Codecov Report

❌ Patch coverage is 74.19355% with 72 lines in your changes missing coverage. Please review.
✅ Project coverage is 81.07%. Comparing base (db8c074) to head (ea9f210).

Files with missing lines	Patch %	Lines
hugr-core/src/hugr/views/syn_edge.rs	31.76%	56 Missing and 2 partials ⚠️
hugr-core/src/hugr/views/sibling_subgraph.rs	90.90%	7 Missing and 1 partial ⚠️
hugr-core/src/hugr/views.rs	92.85%	3 Missing ⚠️
...ugr-core/src/hugr/views/sibling_subgraph/convex.rs	94.33%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3015      +/-   ##
==========================================
- Coverage   81.32%   81.07%   -0.26%     
==========================================
  Files         240      242       +2     
  Lines       45407    45183     -224     
  Branches    39175    38951     -224     
==========================================
- Hits        36929    36633     -296     
- Misses       6488     6585      +97     
+ Partials     1990     1965      -25

Flag	Coverage Δ
python	`88.89% <ø> (ø)`
rust	`79.82% <74.19%> (-0.30%)`	⬇️

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.

…able

…ortgraph left

This reverts commit 04178e4.

codspeed-hq · 2026-04-09T08:52:37Z

Merging this PR will degrade performance by 23.4%

❌ 6 regressed benchmarks
✅ 23 untouched benchmarks
⏩ 6 skipped benchmarks¹

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	`BASE`	`HEAD`	Efficiency
❌	`fewnode_subgraph[10]`	137.5 µs	156 µs	-11.88%
❌	`fewnode_subgraph[1000]`	5.3 ms	7 ms	-23.4%
❌	`multinode_subgraph[1000]`	14.3 ms	15.9 ms	-10.05%
❌	`multinode_subgraph[10]`	135 µs	154.6 µs	-12.65%
❌	`fewnode_subgraph[100]`	613.3 µs	788.4 µs	-22.21%
❌	`multinode_subgraph[100]`	1.1 ms	1.2 ms	-14.15%

_{Comparing acl/syn_edge_prep (ea9f210) with main (db8c074)}

6 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

acl-cqc · 2026-04-09T08:55:00Z

+};
+use std::collections::BTreeSet;
+
+use portgraph::{SecondaryMap, UnmanagedDenseMap};


This file is basically a straight copy of TopoConvexChecker from portgraph then slightly reparametrized

acl-cqc · 2026-04-09T08:55:29Z

@@ -0,0 +1,778 @@
+use std::collections::BTreeSet;


Tests moved out of sibling_subgraph.rs

acl-cqc · 2026-04-09T08:57:41Z

+
+impl<T: LinkView> pv::GraphBase for SynEdgeWrapper<T> {
+    type NodeId = NIdx<T::NodeIndexBase>;
+    type EdgeId = (


Note this is significantly more informative than portgraph's petgraph-traits, but perhaps it shouldn't be

acl-cqc · 2026-04-09T08:57:52Z

+
+impl<T: LinkView> pv::NodeIndexable for SynEdgeWrapper<T> {
+    fn node_bound(&self) -> usize {
+        // ALAN copied from petgraph; but are `NodeIndex`es always dense?


I was a bit worried here??

Ouch, that looks wrong.

It should be self.region_view.node_capacity() instead, since node indices do not get compacted after a removal.

acl-cqc · 2026-04-09T08:58:24Z

+    /// but at least this distinguishes synthetic edges from original edges.
+    ///
+    /// [FlatRegion]: portgraph::view::FlatRegion
+    type EdgeWeight = MaybeSynEdge<(PortOffset<T::PortOffsetBase>, PortOffset<T::PortOffsetBase>)>;


Again, portgraph's petgraph-traits does not provide any edge info, not even just the port offsets

acl-cqc · 2026-04-09T09:02:07Z

+    /// Access to the graph, sufficient to allow [pv::Topo]
+    pub fn petgraph(
+        &self,
+    ) -> impl pv::NodeCount


this type is part of the hiding, and quite deliberate. It could be much more detailed (NodeId/EdgeId types etc.)

acl-cqc · 2026-04-09T09:04:25Z

-    /// Skip convexity check.
-    SkipConvexity,
+mod hidden {
+    #![expect(deprecated)] // Remove enum along with TopoConvexChecker


This mod hidden is so we can do this to capture the deprecation warning from #[derive(Default)]

You can avoid the module by using allow instead of expect. (Seems like a bug, but this works)

#[deprecated] #[allow(deprecated)] #[derive(Default)] pub enum E { #[default] A }

Hah! Yes, that is odd, I had tried expect. Thanks!

acl-cqc · 2026-04-09T09:06:17Z

 /// particularly when many checks must be performed.
-pub type LineConvexChecker<'g, Base> =
-    ConvexChecker<'g, Base, portgraph::algorithms::LineConvexChecker<CheckerRegion<'g, Base>>>;
+pub type LineConvexChecker<'g, Base> = PortgraphCheckerWithNodes<


So I've put some effort into trying to keep this alive, on the grounds that it's potentially much more efficient than TopoConvexChecker. I think it works only for graphs with no container nodes, in which case there will be no syn edges, so it'll still be as ok as it ever was.

But, we could also drop it and plan to reimplement line checking somewhere else if we want it, possibly based off Quantinuum/tket2#1289 somehow

Nice! If it works with your changes I would keep it.

acl-cqc · 2026-04-09T09:08:55Z

+        outputs: &OutgoingPorts<H::Node>,
+        function_calls: &IncomingPorts<H::Node>,
+    ) -> Result<Vec<H::Node>, InvalidSubgraph<H::Node>> {
+        // Compute the nodes inside the boundary ignoring synthetic edges -


This is key, can you check that you agree?

The logic looks good to me

acl-cqc · 2026-04-09T09:09:35Z

+        // that would lead to needing those synthetic edges) then the subgraph is invalid
+        // (nodes would not share the same parent)!
+        let node_indices = make_pg_subgraph::<H>(
+            self.checker.graph().region_view.clone(),


So I guess this is the same backdoor as portgraph_no_syn_edges, and a slight shame I can't use the same method, but here I think it is the right thing to do

aborgna-q

LGTM asides from the node bound bug.

aborgna-q · 2026-04-27T15:57:45Z

-    /// Skip convexity check.
-    SkipConvexity,
+mod hidden {
+    #![expect(deprecated)] // Remove enum along with TopoConvexChecker


You can avoid the module by using allow instead of expect. (Seems like a bug, but this works)

#[deprecated] #[allow(deprecated)] #[derive(Default)] pub enum E { #[default] A }

aborgna-q · 2026-04-27T15:58:38Z

 /// particularly when many checks must be performed.
-pub type LineConvexChecker<'g, Base> =
-    ConvexChecker<'g, Base, portgraph::algorithms::LineConvexChecker<CheckerRegion<'g, Base>>>;
+pub type LineConvexChecker<'g, Base> = PortgraphCheckerWithNodes<


Nice! If it works with your changes I would keep it.

aborgna-q · 2026-04-27T16:01:25Z

+        outputs: &OutgoingPorts<H::Node>,
+        function_calls: &IncomingPorts<H::Node>,
+    ) -> Result<Vec<H::Node>, InvalidSubgraph<H::Node>> {
+        // Compute the nodes inside the boundary ignoring synthetic edges -


The logic looks good to me

aborgna-q · 2026-04-27T16:19:10Z


+    /// A view of a flat region, including ordering constraints from nonlocal edges,
+    /// suitable for use with petgraph algorithms.
+    fn scheduling_graph(&self, parent: Self::Node) -> SchedulingGraph<'_, Self> {


I like scheduling_graph, it conveys well the causal order meaning.

aborgna-q · 2026-05-01T13:45:33Z

+
+impl<T: LinkView> pv::NodeIndexable for SynEdgeWrapper<T> {
+    fn node_bound(&self) -> usize {
+        // ALAN copied from petgraph; but are `NodeIndex`es always dense?


Ouch, that looks wrong.

It should be self.region_view.node_capacity() instead, since node indices do not get compacted after a removal.

acl-cqc added 11 commits March 31, 2026 18:10

Add (local to views) mod syn_edge with SynEdgeWrapper

e010488

Remove commented-out GetAdjacencyMatrix

69eea12

Move SiblingSubgraph tests into new file

b9f83a5

add petgraph convexity checker

12a3fd5

Define scheduling_graph in HugrView

3fcbbdc

Add small test that implements expected traits

4d1d0c8

SchedulingGraph: rename node(s=>_map), add fn graph() -> impl... for …

b813360

…Topo

Deprecate region_portgraph, switch all uses except sibling_subgraph (…

87f0c36

…woot!)

Fix TopoConvexChecker where, add test that we can use on Scheduling…

d532d3c

…Graph

WIP add SchedGraphChecker and try to impl HugrConvexChecker for it

514e23e

WIP2 any closer?

d767be2

acl-cqc added this to the hugr-rs 0.27.X milestone Apr 2, 2026

acl-cqc mentioned this pull request Apr 2, 2026

feat!: do not require an Order edge for each nonlocal Ext edge; compute on demand in scheduling_graph() #2951

Open

acl-cqc force-pushed the acl/syn_edge_prep branch from 1356b19 to 6db3845 Compare April 5, 2026 09:39

acl-cqc added 8 commits April 9, 2026 09:46

Reimplement petgraph traits conditioning on T: LinkView, except Visit…

fbc676e

…able

More tests of what compiles

04178e4

fix sibling_subgraph??

3818cfe

and generalize TopoConvexChecker

9cf4da8

Deprecate stuff in sibling_subgraph, move over...some tricky region_p…

242193e

…ortgraph left

update tests to avoid deprecation

ff2e0cc

Revert "More tests of what compiles"

ae1bf67

This reverts commit 04178e4.

Rename SchedulingGraph::graph -> ::petgraph

521319f

acl-cqc force-pushed the acl/syn_edge_prep branch from 6db3845 to 521319f Compare April 9, 2026 08:49

acl-cqc requested a review from aborgna-q April 9, 2026 08:49

acl-cqc added 3 commits April 9, 2026 09:53

better-backdoors

aadc6b2

syn_edge.rs module comment + remove compilation tests

4137aff

sib_sub: ConvexChecker deprecation should name PortgraphCheckerWithNodes

7525ac4

acl-cqc commented Apr 9, 2026

View reviewed changes

acl-cqc added 2 commits April 9, 2026 10:19

clippy hugr-persistent

e1b6fca

doc fix and doc workaround

db89b21

acl-cqc marked this pull request as ready for review April 9, 2026 09:26

acl-cqc requested a review from a team as a code owner April 9, 2026 09:26

acl-cqc added 9 commits April 9, 2026 17:28

better backdoor

956f087

Remove node_map, expose node_to_pg / pg_to_node

9ad9d74

refactor: inline make_boundary (only called once)

20d2d95

test

d14bf1b

views/tests.rs: add syn-edge test

706ea1c

fmt

f7150f1

Update as_petgraph deprecation note

29ed2fb

also deprecate into_region_portgraph

3d42940

fmt

52a9dd9

aborgna-q approved these changes May 1, 2026

View reviewed changes

acl-cqc added 3 commits May 5, 2026 10:07

correct node_count -> capacity

a210440

allow not expect deprecated works w/out mod hidden

16bbe5f

Merge remote-tracking branch 'origin/main' into acl/syn_edge_prep

ea9f210

acl-cqc added this pull request to the merge queue May 5, 2026

Merged via the queue into main with commit d6e11c8 May 5, 2026
28 of 30 checks passed

acl-cqc deleted the acl/syn_edge_prep branch May 5, 2026 09:50

This was referenced Apr 30, 2026

chore: release 0.27.1 #3014

Merged

chore: release #3048

Draft

acl-cqc mentioned this pull request May 5, 2026

feat: upgrade to hugr v0.27.1 Quantinuum/tket2#1568

Merged

Conversation

acl-cqc commented Apr 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov Bot commented Apr 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

codspeed-hq Bot commented Apr 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Merging this PR will degrade performance by 23.4%

Performance Changes

Footnotes

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

acl-cqc Apr 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

aborgna-q left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

acl-cqc commented Apr 2, 2026 •

edited

Loading

codecov Bot commented Apr 5, 2026 •

edited

Loading

codspeed-hq Bot commented Apr 9, 2026 •

edited

Loading

acl-cqc Apr 9, 2026 •

edited

Loading