Repository navigation
fix: missing-asset correctness follow-ups from #10856 - #11233
Conversation
1. ErrorOverlay count reflected node-selection filtering. useErrorGroups exposed filteredMissingModelGroups under the name missingModelGroups. ErrorOverlay read that alias for its model count label, so selecting a node shrunk the overlay's total. Expose both unfiltered (missingModelGroups / missingMediaGroups) and filtered variants so ErrorOverlay gets the true total and the errors tab still gets selection-scoped groups. TabErrors renamed its destructure. 2. bypass/un-bypass of a missing-model node dropped url/hash metadata. Realtime scanNodeModelCandidates only reads widget values, so un-bypass produced a fresh candidate without the url that enrichWithEmbeddedMetadata had previously attached from graphData.models. Add enrichCandidateFromNodeProperties which copies url/hash/directory from the node's own properties.models (persistent across mode toggles). Applies to every call site of the per-node scan, so paste, un-bypass, and mode change all restore the metadata. 3. Initial full scan surfaced interior errors of muted/bypassed subgraph containers. scanAllModelCandidates, scanAllMediaCandidates, and the JSON missing-node scan only check each node's own mode; interior nodes whose parent container is bypassed passed through. Runtime handleNodeModeChange already filters correctly — the bug was scoped to the initial-load path. Add isAncestorPathActive to graphTraversalUtil and post-filter the three pipelines in runMissingModelPipeline / runMissingMediaPipeline / loadGraphData before the results are cached or surfaced.
The ancestor-prefix iteration in isAncestorPathActive reimplemented what getParentExecutionIds (src/types/nodeIdentification.ts) already provides. Use the shared helper — same semantics, less manual string work, and consistent with how missingModel/Media/Nodes stores already build ancestor prefixes.
useGraphNodeManager replays graph.onNodeAdded(node) for each existing interior node when the Vue node manager initializes on subgraph entry. That chain reaches scanSingleNodeErrors via installErrorClearingHooks' onNodeAdded override. Each interior node's own mode is active, so the caller guards pass and the scan ran — reintroducing missing-asset errors that the initial pipeline's ancestor post-filter had correctly dropped. Add an ancestor-activity gate at the top of scanSingleNodeErrors, the single entry point shared by paste, un-bypass, subgraph entry, and subgraph container activation. One guard covers every realtime scan path.
Adds end-to-end regression tests for each bug addressed in this branch: 1. ErrorOverlay model count must not reflect node selection (errorOverlay.spec.ts) — new fixture missing_models_distinct.json with two nodes referencing two different model names. Selecting a node must not change the "N required models are missing" label. 2. Bypass/un-bypass cycle preserves Copy URL button (errorsTabModeAware.spec.ts) — uses missing_models_from_node_properties.json because the fix restores url/hash/directory from node.properties.models on realtime scan; workflow-level models[] is not seen by that path. 3. Initial load with bypassed subgraph container does not surface interior errors (errorsTabModeAware.spec.ts) — new fixture missing_models_in_bypassed_subgraph.json (subgraph container mode=BYPASS, interior CheckpointLoader active, unrelated KSampler active). 4. Entering a bypassed subgraph does not resurface interior errors (errorsTabModeAware.spec.ts) — shares the fixture from (3); navigates into the bypassed subgraph and verifies the errors tab stays hidden.
Three narrowly-scoped defensive fixes found during parallel subagent review of this branch: 1. enrichCandidateFromNodeProperties now requires directory agreement when the candidate already carries one. Without this, a node referencing the same model name under two directories (e.g. a LoRA present in multiple folders) could get the wrong url/hash stamped onto a candidate — name-only matching was too loose. Mirrors the directory check already present in enrichWithEmbeddedMetadata. 2. scanSingleNodeErrors now skips when its execution ID cannot be resolved (node is detached or has no graph). The previous guard short-circuited only on a falsy rootGraph check, so a detached node whose own mode looked active could still drive the missing- node-type scan path. 3. isCandidateStillActive in the realtime verification callback now also confirms the ancestor path is active, mirroring the initial-pipeline post-filter. Without this, re-bypassing an outer subgraph during an in-flight per-node verification would leak a stale missing-asset entry into the store. Added one unit test for the directory-collision path.
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
📝 WalkthroughWalkthroughAdds ancestor-path activation checks to suppress missing-model/media errors inside bypassed/NEVER subgraph containers, enriches missing-model candidates from node Changes
Sequence DiagramssequenceDiagram
participant UI as Error UI
participant App as App Pipeline
participant Scanner as Node Scanner
participant Traversal as Graph Traversal
participant Enricher as Candidate Enricher
participant Store as MissingModel Store
UI->>App: loadGraphData / runMissingModelPipeline
App->>Scanner: scan nodes -> produce candidates
Scanner->>Traversal: isAncestorPathActive(rootGraph, execId)
alt Ancestor path inactive
Traversal-->>Scanner: inactive
Scanner-->>App: drop candidate (do not enrich/verify)
else Ancestor path active
Traversal-->>Scanner: active
Scanner->>Enricher: enrichCandidateFromNodeProperties(candidate, embeddedModels)
Enricher-->>App: return enriched candidate
App->>Store: verify/add candidate (async cloud check)
Store->>Traversal: isMissingCandidateActive(rootGraph, candidate) (post-verify)
alt still active
Store-->>UI: surface missing-model entry
else no longer active
Store-->>App: ignore stale verification result
end
end
UI->>UI: group errors (unfiltered vs selection-filtered)
Estimated code review effort🎯 4 (Complex) | ⏱️ ~50 minutes Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
🎨 Storybook: ✅ Built — View Storybook |
🎭 Playwright: ✅ 1120 passed, 0 failed · 6 flaky📊 Browser Reports
|
📦 Bundle: 5.16 MB gzip 🔴 +1.21 kBDetailsSummary
Category Glance App Entry Points — 22.3 kB (baseline 22.3 kB) • ⚪ 0 BMain entry bundles and manifests
Status: 1 added / 1 removed Graph Workspace — 1.22 MB (baseline 1.22 MB) • 🔴 +482 BGraph editor runtime, canvas, workflow orchestration
Status: 1 added / 1 removed Views & Navigation — 76.6 kB (baseline 76.6 kB) • ⚪ 0 BTop-level views, pages, and routed surfaces
Status: 9 added / 9 removed / 2 unchanged Panels & Settings — 482 kB (baseline 482 kB) • ⚪ 0 BConfiguration panels, inspectors, and settings screens
Status: 10 added / 10 removed / 11 unchanged User & Accounts — 17.1 kB (baseline 17.1 kB) • ⚪ 0 BAuthentication, profile, and account management bundles
Status: 5 added / 5 removed / 2 unchanged Editors & Dialogs — 109 kB (baseline 109 kB) • ⚪ 0 BModals, dialogs, drawers, and in-app editors
Status: 2 added / 2 removed UI Components — 60.3 kB (baseline 60.3 kB) • ⚪ 0 BReusable component library chunks
Status: 5 added / 5 removed / 8 unchanged Data & Services — 3 MB (baseline 3 MB) • 🔴 +3.73 kBStores, services, APIs, and repositories
Status: 13 added / 13 removed / 4 unchanged Utilities & Hooks — 344 kB (baseline 344 kB) • ⚪ 0 BHelpers, composables, and utility bundles
Status: 13 added / 13 removed / 14 unchanged Vendor & Third-Party — 9.86 MB (baseline 9.86 MB) • ⚪ 0 BExternal libraries and shared vendor chunks Status: 16 unchanged Other — 8.57 MB (baseline 8.57 MB) • ⚪ 0 BBundles that do not match a named category
Status: 56 added / 56 removed / 79 unchanged ⚡ Performance Report
No regressions detected. All metrics
Historical variance (last 15 runs)
Trend (last 15 commits on main)
Raw data{
"timestamp": "2026-04-15T09:14:19.920Z",
"gitSha": "0cfc359108d116ee4104e2a43b4afc92dcff7f78",
"branch": "fix/missing-asset-followups-from-10856",
"measurements": [
{
"name": "canvas-idle",
"durationMs": 2032.496000000009,
"styleRecalcs": 12,
"styleRecalcDurationMs": 10.623,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 356.557,
"heapDeltaBytes": 20388092,
"heapUsedBytes": 63206952,
"domNodes": 24,
"jsHeapTotalBytes": 22544384,
"scriptDurationMs": 27.726,
"eventListeners": 10,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "canvas-idle",
"durationMs": 2010.2570000000242,
"styleRecalcs": 9,
"styleRecalcDurationMs": 7.9940000000000015,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 333.557,
"heapDeltaBytes": 20030464,
"heapUsedBytes": 62941620,
"domNodes": 18,
"jsHeapTotalBytes": 22544384,
"scriptDurationMs": 15.653999999999998,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.699999999999818
},
{
"name": "canvas-idle",
"durationMs": 2024.3010000000368,
"styleRecalcs": 11,
"styleRecalcDurationMs": 9.959,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 355.91100000000006,
"heapDeltaBytes": -4610444,
"heapUsedBytes": 44342548,
"domNodes": 22,
"jsHeapTotalBytes": 24903680,
"scriptDurationMs": 23.345000000000006,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "canvas-mouse-sweep",
"durationMs": 2043.2720000000018,
"styleRecalcs": 80,
"styleRecalcDurationMs": 45.69,
"layouts": 12,
"layoutDurationMs": 3.3720000000000003,
"taskDurationMs": 981.066,
"heapDeltaBytes": 15984480,
"heapUsedBytes": 58915416,
"domNodes": 66,
"jsHeapTotalBytes": 22806528,
"scriptDurationMs": 144.62,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333335,
"p95FrameDurationMs": 16.699999999999818
},
{
"name": "canvas-mouse-sweep",
"durationMs": 2041.4349999999786,
"styleRecalcs": 83,
"styleRecalcDurationMs": 44.743,
"layouts": 12,
"layoutDurationMs": 3.62,
"taskDurationMs": 964.982,
"heapDeltaBytes": 16314192,
"heapUsedBytes": 59182908,
"domNodes": 69,
"jsHeapTotalBytes": 23068672,
"scriptDurationMs": 134.947,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.699999999999818
},
{
"name": "canvas-mouse-sweep",
"durationMs": 2031.5919999999323,
"styleRecalcs": 85,
"styleRecalcDurationMs": 48.361000000000004,
"layouts": 12,
"layoutDurationMs": 3.539,
"taskDurationMs": 992.836,
"heapDeltaBytes": 16274016,
"heapUsedBytes": 59030356,
"domNodes": 68,
"jsHeapTotalBytes": 23330816,
"scriptDurationMs": 139.98,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333335,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "canvas-zoom-sweep",
"durationMs": 1707.5609999999983,
"styleRecalcs": 31,
"styleRecalcDurationMs": 17.360999999999997,
"layouts": 6,
"layoutDurationMs": 0.667,
"taskDurationMs": 290.62,
"heapDeltaBytes": 15365240,
"heapUsedBytes": 67092436,
"domNodes": 79,
"jsHeapTotalBytes": 22806528,
"scriptDurationMs": 23.292,
"eventListeners": 19,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "canvas-zoom-sweep",
"durationMs": 1743.5560000000123,
"styleRecalcs": 32,
"styleRecalcDurationMs": 17.468,
"layouts": 6,
"layoutDurationMs": 0.6569999999999999,
"taskDurationMs": 300.89700000000005,
"heapDeltaBytes": 15712108,
"heapUsedBytes": 67284480,
"domNodes": 82,
"jsHeapTotalBytes": 23068672,
"scriptDurationMs": 25.212,
"eventListeners": 19,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "canvas-zoom-sweep",
"durationMs": 1731.1849999999822,
"styleRecalcs": 32,
"styleRecalcDurationMs": 18.392,
"layouts": 6,
"layoutDurationMs": 0.702,
"taskDurationMs": 295.891,
"heapDeltaBytes": 24698920,
"heapUsedBytes": 67377892,
"domNodes": 80,
"jsHeapTotalBytes": 20447232,
"scriptDurationMs": 24.02,
"eventListeners": 19,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333335,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "dom-widget-clipping",
"durationMs": 588.577000000015,
"styleRecalcs": 12,
"styleRecalcDurationMs": 10.041999999999998,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 376.17300000000006,
"heapDeltaBytes": 6343696,
"heapUsedBytes": 49481020,
"domNodes": 20,
"jsHeapTotalBytes": 13369344,
"scriptDurationMs": 67.80600000000001,
"eventListeners": 2,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "dom-widget-clipping",
"durationMs": 568.0920000000356,
"styleRecalcs": 14,
"styleRecalcDurationMs": 9.992,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 344.677,
"heapDeltaBytes": 6892528,
"heapUsedBytes": 49468016,
"domNodes": 23,
"jsHeapTotalBytes": 12845056,
"scriptDurationMs": 65.897,
"eventListeners": 2,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333332,
"p95FrameDurationMs": 16.700000000000273
},
{
"name": "dom-widget-clipping",
"durationMs": 568.0879999999888,
"styleRecalcs": 12,
"styleRecalcDurationMs": 12.107000000000001,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 342.245,
"heapDeltaBytes": 6448240,
"heapUsedBytes": 49439896,
"domNodes": 20,
"jsHeapTotalBytes": 13631488,
"scriptDurationMs": 68.263,
"eventListeners": 2,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.663333333333338,
"p95FrameDurationMs": 16.700000000000273
},
{
"name": "large-graph-idle",
"durationMs": 2043.6010000000238,
"styleRecalcs": 10,
"styleRecalcDurationMs": 9.930000000000001,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 531.868,
"heapDeltaBytes": 4701584,
"heapUsedBytes": 55629940,
"domNodes": -258,
"jsHeapTotalBytes": 16703488,
"scriptDurationMs": 97.6,
"eventListeners": -125,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.699999999999818
},
{
"name": "large-graph-idle",
"durationMs": 2057.1710000000394,
"styleRecalcs": 10,
"styleRecalcDurationMs": 9.131999999999998,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 526.848,
"heapDeltaBytes": 4678652,
"heapUsedBytes": 55985576,
"domNodes": -259,
"jsHeapTotalBytes": 16285696,
"scriptDurationMs": 105.75899999999999,
"eventListeners": -125,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "large-graph-idle",
"durationMs": 2051.4249999999947,
"styleRecalcs": 12,
"styleRecalcDurationMs": 10.912999999999998,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 533.7019999999999,
"heapDeltaBytes": 5183484,
"heapUsedBytes": 55974948,
"domNodes": -254,
"jsHeapTotalBytes": 16965632,
"scriptDurationMs": 98.594,
"eventListeners": -125,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "large-graph-pan",
"durationMs": 2137.8890000000297,
"styleRecalcs": 69,
"styleRecalcDurationMs": 15.145,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 1054.964,
"heapDeltaBytes": 19093292,
"heapUsedBytes": 72693128,
"domNodes": -259,
"jsHeapTotalBytes": 17432576,
"scriptDurationMs": 398.367,
"eventListeners": -127,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "large-graph-pan",
"durationMs": 2138.4879999999953,
"styleRecalcs": 69,
"styleRecalcDurationMs": 15.671000000000005,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 1060.62,
"heapDeltaBytes": 16569696,
"heapUsedBytes": 68817248,
"domNodes": -260,
"jsHeapTotalBytes": 18743296,
"scriptDurationMs": 396.94399999999996,
"eventListeners": -127,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.699999999999818
},
{
"name": "large-graph-pan",
"durationMs": 2137.5759999999673,
"styleRecalcs": 70,
"styleRecalcDurationMs": 15.98,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 1051.9820000000002,
"heapDeltaBytes": 259272,
"heapUsedBytes": 53752736,
"domNodes": -259,
"jsHeapTotalBytes": 18219008,
"scriptDurationMs": 389.43000000000006,
"eventListeners": -123,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "large-graph-zoom",
"durationMs": 3144.0880000000107,
"styleRecalcs": 66,
"styleRecalcDurationMs": 16.052999999999997,
"layouts": 60,
"layoutDurationMs": 7.123000000000001,
"taskDurationMs": 1266.792,
"heapDeltaBytes": 7425296,
"heapUsedBytes": 62032472,
"domNodes": -262,
"jsHeapTotalBytes": 16441344,
"scriptDurationMs": 480.25800000000004,
"eventListeners": -123,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "large-graph-zoom",
"durationMs": 3148.1860000000097,
"styleRecalcs": 66,
"styleRecalcDurationMs": 16.087999999999997,
"layouts": 60,
"layoutDurationMs": 7.101,
"taskDurationMs": 1276.6219999999998,
"heapDeltaBytes": 6991452,
"heapUsedBytes": 62275940,
"domNodes": -264,
"jsHeapTotalBytes": 16965632,
"scriptDurationMs": 479.78999999999996,
"eventListeners": -125,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333332,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "large-graph-zoom",
"durationMs": 3149.1640000000416,
"styleRecalcs": 65,
"styleRecalcDurationMs": 15.303999999999998,
"layouts": 60,
"layoutDurationMs": 7.028999999999999,
"taskDurationMs": 1273.271,
"heapDeltaBytes": 6996560,
"heapUsedBytes": 61635200,
"domNodes": -267,
"jsHeapTotalBytes": 16965632,
"scriptDurationMs": 480.66400000000004,
"eventListeners": -125,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "minimap-idle",
"durationMs": 2051.7189999999914,
"styleRecalcs": 11,
"styleRecalcDurationMs": 9.784999999999998,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 514.488,
"heapDeltaBytes": 4334536,
"heapUsedBytes": 57206616,
"domNodes": -259,
"jsHeapTotalBytes": 16441344,
"scriptDurationMs": 95.586,
"eventListeners": -127,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333332,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "minimap-idle",
"durationMs": 2013.388999999961,
"styleRecalcs": 8,
"styleRecalcDurationMs": 7.615,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 506.06700000000006,
"heapDeltaBytes": 2534244,
"heapUsedBytes": 56856516,
"domNodes": -263,
"jsHeapTotalBytes": 16965632,
"scriptDurationMs": 87.625,
"eventListeners": -127,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333332,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "minimap-idle",
"durationMs": 1993.488999999954,
"styleRecalcs": 9,
"styleRecalcDurationMs": 7.962000000000001,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 506.28799999999995,
"heapDeltaBytes": 16612584,
"heapUsedBytes": 70363980,
"domNodes": -259,
"jsHeapTotalBytes": 14925824,
"scriptDurationMs": 89.353,
"eventListeners": -127,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.699999999999818
},
{
"name": "subgraph-dom-widget-clipping",
"durationMs": 535.6150000000071,
"styleRecalcs": 48,
"styleRecalcDurationMs": 11.932,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 345.31899999999996,
"heapDeltaBytes": 6816676,
"heapUsedBytes": 49655692,
"domNodes": 22,
"jsHeapTotalBytes": 13107200,
"scriptDurationMs": 121.919,
"eventListeners": 8,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000273
},
{
"name": "subgraph-dom-widget-clipping",
"durationMs": 529.9260000000459,
"styleRecalcs": 47,
"styleRecalcDurationMs": 10.965,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 343.001,
"heapDeltaBytes": 6552492,
"heapUsedBytes": 49720712,
"domNodes": 20,
"jsHeapTotalBytes": 13369344,
"scriptDurationMs": 125.171,
"eventListeners": 8,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.663333333333338,
"p95FrameDurationMs": 16.700000000000273
},
{
"name": "subgraph-dom-widget-clipping",
"durationMs": 521.7830000000276,
"styleRecalcs": 47,
"styleRecalcDurationMs": 11.300999999999998,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 344.69500000000005,
"heapDeltaBytes": 6413108,
"heapUsedBytes": 49529696,
"domNodes": 20,
"jsHeapTotalBytes": 13631488,
"scriptDurationMs": 119.726,
"eventListeners": 8,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333332,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "subgraph-idle",
"durationMs": 2013.1000000000085,
"styleRecalcs": 10,
"styleRecalcDurationMs": 8.592000000000002,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 321.082,
"heapDeltaBytes": 19830572,
"heapUsedBytes": 63057580,
"domNodes": 20,
"jsHeapTotalBytes": 23330816,
"scriptDurationMs": 13.977000000000004,
"eventListeners": 4,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "subgraph-idle",
"durationMs": 2014.521000000002,
"styleRecalcs": 10,
"styleRecalcDurationMs": 8.767999999999999,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 334.01399999999995,
"heapDeltaBytes": 19911140,
"heapUsedBytes": 63103092,
"domNodes": 20,
"jsHeapTotalBytes": 23068672,
"scriptDurationMs": 18.394000000000002,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "subgraph-idle",
"durationMs": 2020.483000000013,
"styleRecalcs": 11,
"styleRecalcDurationMs": 10.802,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 334.57300000000004,
"heapDeltaBytes": 19953628,
"heapUsedBytes": 63105600,
"domNodes": 22,
"jsHeapTotalBytes": 22806528,
"scriptDurationMs": 18.148,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "subgraph-mouse-sweep",
"durationMs": 1974.1659999999968,
"styleRecalcs": 84,
"styleRecalcDurationMs": 49.160999999999994,
"layouts": 16,
"layoutDurationMs": 4.686,
"taskDurationMs": 892.894,
"heapDeltaBytes": 11818876,
"heapUsedBytes": 55039328,
"domNodes": 73,
"jsHeapTotalBytes": 23068672,
"scriptDurationMs": 97.676,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333332,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "subgraph-mouse-sweep",
"durationMs": 1984.0599999999995,
"styleRecalcs": 88,
"styleRecalcDurationMs": 45.836999999999996,
"layouts": 16,
"layoutDurationMs": 4.9830000000000005,
"taskDurationMs": 896.5609999999999,
"heapDeltaBytes": 11750312,
"heapUsedBytes": 55140892,
"domNodes": 74,
"jsHeapTotalBytes": 23068672,
"scriptDurationMs": 96.208,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.66333333333335,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "subgraph-mouse-sweep",
"durationMs": 1978.0499999999392,
"styleRecalcs": 87,
"styleRecalcDurationMs": 43.173,
"layouts": 16,
"layoutDurationMs": 4.464,
"taskDurationMs": 916.4939999999999,
"heapDeltaBytes": 2793952,
"heapUsedBytes": 54692308,
"domNodes": 74,
"jsHeapTotalBytes": 25690112,
"scriptDurationMs": 96.943,
"eventListeners": 6,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.800000000000182
},
{
"name": "viewport-pan-sweep",
"durationMs": 8170.3779999999715,
"styleRecalcs": 253,
"styleRecalcDurationMs": 44.824999999999996,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 3709.4809999999998,
"heapDeltaBytes": 31747964,
"heapUsedBytes": 83052888,
"domNodes": -251,
"jsHeapTotalBytes": 20606976,
"scriptDurationMs": 1320.221,
"eventListeners": -105,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "viewport-pan-sweep",
"durationMs": 8186.0960000000205,
"styleRecalcs": 250,
"styleRecalcDurationMs": 43.223,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 3560.428,
"heapDeltaBytes": 24104632,
"heapUsedBytes": 76929960,
"domNodes": -260,
"jsHeapTotalBytes": 18743296,
"scriptDurationMs": 1232.2810000000002,
"eventListeners": -111,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.80000000000109
},
{
"name": "viewport-pan-sweep",
"durationMs": 8158.2389999999805,
"styleRecalcs": 252,
"styleRecalcDurationMs": 43.85600000000001,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 3567.9399999999996,
"heapDeltaBytes": 22269152,
"heapUsedBytes": 73269144,
"domNodes": -256,
"jsHeapTotalBytes": 19791872,
"scriptDurationMs": 1239.2510000000002,
"eventListeners": -111,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "vue-large-graph-idle",
"durationMs": 12446.489999999983,
"styleRecalcs": 1,
"styleRecalcDurationMs": 18.884999999999984,
"layouts": 1,
"layoutDurationMs": 4.779999999999992,
"taskDurationMs": 12407.073,
"heapDeltaBytes": -31814792,
"heapUsedBytes": 161478336,
"domNodes": -8331,
"jsHeapTotalBytes": 27877376,
"scriptDurationMs": 601.228,
"eventListeners": -14698,
"totalBlockingTimeMs": 1243,
"frameDurationMs": 17.219999999999953,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "vue-large-graph-idle",
"durationMs": 12155.105000000049,
"styleRecalcs": 1,
"styleRecalcDurationMs": 16.969000000000012,
"layouts": 1,
"layoutDurationMs": 4.869999999999999,
"taskDurationMs": 12131.372000000001,
"heapDeltaBytes": -16979184,
"heapUsedBytes": 168520288,
"domNodes": -8331,
"jsHeapTotalBytes": 36921344,
"scriptDurationMs": 607.004,
"eventListeners": -14700,
"totalBlockingTimeMs": 1199,
"frameDurationMs": 17.219999999999953,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "vue-large-graph-idle",
"durationMs": 12006.755999999996,
"styleRecalcs": 1,
"styleRecalcDurationMs": 16.50199999999999,
"layouts": 1,
"layoutDurationMs": 4.759000000000013,
"taskDurationMs": 11984.160999999998,
"heapDeltaBytes": -47009404,
"heapUsedBytes": 160220764,
"domNodes": -8329,
"jsHeapTotalBytes": 28925952,
"scriptDurationMs": 571.191,
"eventListeners": -14702,
"totalBlockingTimeMs": 1203,
"frameDurationMs": 17.223333333333358,
"p95FrameDurationMs": 16.700000000000728
},
{
"name": "vue-large-graph-pan",
"durationMs": 14195.175000000007,
"styleRecalcs": 66,
"styleRecalcDurationMs": 12.778999999999986,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 14158.171,
"heapDeltaBytes": -54612172,
"heapUsedBytes": 167106460,
"domNodes": -9850,
"jsHeapTotalBytes": 24993792,
"scriptDurationMs": 883.9449999999999,
"eventListeners": -23955,
"totalBlockingTimeMs": 0,
"frameDurationMs": 17.776666666666642,
"p95FrameDurationMs": 33.29999999999927
},
{
"name": "vue-large-graph-pan",
"durationMs": 13985.25499999994,
"styleRecalcs": 66,
"styleRecalcDurationMs": 12.861000000000011,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 13953.545999999998,
"heapDeltaBytes": -37094780,
"heapUsedBytes": 177387808,
"domNodes": -9848,
"jsHeapTotalBytes": 27791360,
"scriptDurationMs": 860.1410000000001,
"eventListeners": -23957,
"totalBlockingTimeMs": 0,
"frameDurationMs": 17.219999999999953,
"p95FrameDurationMs": 16.799999999999272
},
{
"name": "vue-large-graph-pan",
"durationMs": 14338.303999999993,
"styleRecalcs": 67,
"styleRecalcDurationMs": 13.959,
"layouts": 0,
"layoutDurationMs": 0,
"taskDurationMs": 14317.125,
"heapDeltaBytes": -34898068,
"heapUsedBytes": 162294680,
"domNodes": -9850,
"jsHeapTotalBytes": -11706368,
"scriptDurationMs": 863.602,
"eventListeners": -23953,
"totalBlockingTimeMs": 3,
"frameDurationMs": 17.223333333333358,
"p95FrameDurationMs": 16.80000000000291
},
{
"name": "workflow-execution",
"durationMs": 117.18000000001894,
"styleRecalcs": 10,
"styleRecalcDurationMs": 20.380999999999997,
"layouts": 4,
"layoutDurationMs": 1.827,
"taskDurationMs": 91.17499999999998,
"heapDeltaBytes": 2865072,
"heapUsedBytes": 47484872,
"domNodes": 146,
"jsHeapTotalBytes": 262144,
"scriptDurationMs": 23.91,
"eventListeners": 37,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.663333333333338,
"p95FrameDurationMs": 16.700000000000273
},
{
"name": "workflow-execution",
"durationMs": 439.4369999999981,
"styleRecalcs": 17,
"styleRecalcDurationMs": 21.209000000000003,
"layouts": 5,
"layoutDurationMs": 1.147,
"taskDurationMs": 116.42699999999999,
"heapDeltaBytes": 4474484,
"heapUsedBytes": 48153224,
"domNodes": 158,
"jsHeapTotalBytes": 0,
"scriptDurationMs": 29.544999999999998,
"eventListeners": 71,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.663333333333338,
"p95FrameDurationMs": 16.700000000000273
},
{
"name": "workflow-execution",
"durationMs": 453.68699999994533,
"styleRecalcs": 20,
"styleRecalcDurationMs": 27.179000000000002,
"layouts": 6,
"layoutDurationMs": 1.5439999999999998,
"taskDurationMs": 131.22400000000002,
"heapDeltaBytes": 4547152,
"heapUsedBytes": 55067040,
"domNodes": 158,
"jsHeapTotalBytes": 262144,
"scriptDurationMs": 30.686999999999998,
"eventListeners": 71,
"totalBlockingTimeMs": 0,
"frameDurationMs": 16.666666666666668,
"p95FrameDurationMs": 16.700000000000273
}
]
} |
Codecov Report❌ Patch coverage is
@@ Coverage Diff @@
## main #11233 +/- ##
==========================================
+ Coverage 43.70% 43.81% +0.11%
==========================================
Files 1349 1350 +1
Lines 69152 69273 +121
Branches 19159 19218 +59
==========================================
+ Hits 30224 30354 +130
+ Misses 38330 38320 -10
- Partials 598 599 +1
Flags with carried forward coverage won't be shown. Click here to find out more.
... and 12 files with indirect coverage changes 🚀 New features to boost your workflow:
|
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (2)
src/components/rightSidePanel/errors/useErrorGroups.test.ts (1)
767-780: Selection-independence test doesn’t currently exercise selection state.The test name asserts behavior “regardless of selection,” but no node-selection input is set. Consider explicitly setting a selection and asserting both: unfiltered groups keep the total, while filtered groups narrow.
As per coding guidelines: “Aim for behavioral coverage of critical and new features in tests.”
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/components/rightSidePanel/errors/useErrorGroups.test.ts` around lines 767 - 780, The test missingModelGroups returns total candidates regardless of selection (ErrorOverlay contract) currently doesn't set any selection state; update the test (around createErrorGroups, store.surfaceMissingModels, groups.missingModelGroups) to explicitly set a node selection (e.g., simulate selecting a different node via the store or whatever selection API exists on the test helper) before awaiting nextTick, then assert two things: 1) groups.missingModelGroups still returns the full unfiltered set (length 1 with 2 models) and 2) the filtered view (use the selection-aware accessor or method your code provides—e.g., groups.filteredModelGroups or store.getFilteredGroups) reflects the selection and narrows the list; this ensures the test actually exercises selection-independence of missingModelGroups while validating the selection-based filtering behavior.src/platform/missingModel/missingModelScan.ts (1)
144-145: Consider using a type guard or narrower cast for type safety.The inline cast
(node as { properties?: { models?: ModelFile[] } })works but is somewhat verbose. This pattern is repeated elsewhere, so it's acceptable, but consider extracting a helper if this pattern proliferates.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/platform/missingModel/missingModelScan.ts` around lines 144 - 145, The inline cast for extracting models from node (const embeddedModels = (node as { properties?: { models?: ModelFile[] } }).properties?.models) is verbose and fragile; replace it with a type guard or a small helper function (e.g., isNodeWithModels(node): node is { properties: { models: ModelFile[] } }) and use that guard to safely access properties.models, updating usages in missingModelScan.ts where embeddedModels and node are referenced so you avoid repetitive inline casts and improve type safety.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@browser_tests/assets/missing/missing_models_distinct.json`:
- Around line 9-10: The JSON fixture contains node "size" entries set to [315,
98] which is below the required minimum; update the "size" key values (e.g., the
entry currently [315, 98]) to at least [400, 200] in
missing_models_distinct.json and scan the same file for any other "size" entries
(the other small nodes noted in the diff) and set them to [400, 200] as well to
comply with the browser_tests/assets minimum node size rule.
---
Nitpick comments:
In `@src/components/rightSidePanel/errors/useErrorGroups.test.ts`:
- Around line 767-780: The test missingModelGroups returns total candidates
regardless of selection (ErrorOverlay contract) currently doesn't set any
selection state; update the test (around createErrorGroups,
store.surfaceMissingModels, groups.missingModelGroups) to explicitly set a node
selection (e.g., simulate selecting a different node via the store or whatever
selection API exists on the test helper) before awaiting nextTick, then assert
two things: 1) groups.missingModelGroups still returns the full unfiltered set
(length 1 with 2 models) and 2) the filtered view (use the selection-aware
accessor or method your code provides—e.g., groups.filteredModelGroups or
store.getFilteredGroups) reflects the selection and narrows the list; this
ensures the test actually exercises selection-independence of missingModelGroups
while validating the selection-based filtering behavior.
In `@src/platform/missingModel/missingModelScan.ts`:
- Around line 144-145: The inline cast for extracting models from node (const
embeddedModels = (node as { properties?: { models?: ModelFile[] }
}).properties?.models) is verbose and fragile; replace it with a type guard or a
small helper function (e.g., isNodeWithModels(node): node is { properties: {
models: ModelFile[] } }) and use that guard to safely access properties.models,
updating usages in missingModelScan.ts where embeddedModels and node are
referenced so you avoid repetitive inline casts and improve type safety.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: 081f4c68-b7ed-43f6-b554-3eaa429d7ef0
📒 Files selected for processing (14)
browser_tests/assets/missing/missing_models_distinct.jsonbrowser_tests/assets/missing/missing_models_in_bypassed_subgraph.jsonbrowser_tests/tests/errorOverlay.spec.tsbrowser_tests/tests/propertiesPanel/errorsTabModeAware.spec.tssrc/components/rightSidePanel/errors/TabErrors.vuesrc/components/rightSidePanel/errors/useErrorGroups.test.tssrc/components/rightSidePanel/errors/useErrorGroups.tssrc/composables/graph/useErrorClearingHooks.test.tssrc/composables/graph/useErrorClearingHooks.tssrc/platform/missingModel/missingModelScan.test.tssrc/platform/missingModel/missingModelScan.tssrc/scripts/app.tssrc/utils/graphTraversalUtil.test.tssrc/utils/graphTraversalUtil.ts
The fixture's properties.models[0].directory was 'text_encoders', but the node is a CheckpointLoaderSimple whose modelToNodeStore category is 'checkpoints'. The old name-only enrichment silently tolerated the mismatch, so the test's Copy URL button assertion passed. The new enrichCandidateFromNodeProperties rejects mismatched directories (guards against stamping the wrong url/hash onto a collision candidate), which correctly skipped enrichment and hid the button. Updating the fixture aligns it with the node's real directory and restores the intended coverage.
- missing_models_distinct.json: bump both node sizes from [315, 98] to [400, 200] to follow the browser_tests asset convention (minimum size avoids navigateIntoSubgraph timeout issues downstream). - useErrorGroups.test.ts: the 'returns total regardless of selection' case did not simulate selection, so both unfiltered and filtered variants would have returned the same value and the test could not prove the ErrorOverlay contract. Set canvasStore.selectedItems and mock isLGraphNode to true for the simulated node; assert the unfiltered total stays at 2 and that the filtered shape diverges. CodeRabbit also flagged an inline cast in missingModelScan.ts (enrichCandidateFromNodeProperties) as a nitpick; left as-is since it is a single call site and the reviewer marked it acceptable.
dante01yoon
left a comment
There was a problem hiding this comment.
Two correctness issues still look open.
-
The initial Cloud verification path can still surface stale missing-model/media errors after a node or enclosing subgraph has been bypassed. In
src/scripts/app.ts, the new ancestor filtering happens before async verification starts, but theverifyAssetSupportedCandidates(...).then(...)andverifyCloudMediaCandidates(...).then(...)branches surface confirmed candidates without re-checking whether each candidate is still active at resolution time. The realtime path insrc/composables/graph/useErrorClearingHooks.tsnow does that re-check, so the initial-load path and realtime path are no longer symmetric. -
The workflow-level
graphData.modelsfallback is still ancestor-blind for bypassed subgraphs.isModelReferencedByActiveNode()insrc/platform/missingModel/missingModelScan.tsonly checks the referenced node's ownmode, not whether one of its ancestor subgraph containers is bypassed. When that helper keeps an unmatched workflow-level model alive, the emitted candidate has nonodeId, so the laterisAncestorPathActive(...)filtering insrc/scripts/app.tscannot remove it. A bypassed subgraph containing the only reference can therefore still surface a top-level missing-model row viagraphData.models.
Bug fixes driven by PR review: 1. Cloud verification staleness (app.ts runMissingModelPipeline / runMissingMediaPipeline): the async verify().then() surfaced confirmed candidates without re-checking whether the user had bypassed an enclosing subgraph mid-verification. Both pipelines now apply the ancestor check symmetrically with the realtime path. 2. Workflow-level reference ancestor-blind (missingModelScan.ts isModelReferencedByActiveNode): the helper only checked each referencing node's own mode, ignoring bypassed subgraph containers. A bypassed subgraph whose interior was the sole reference could still surface a top-level missing-model row. The helper now walks ancestors via getParentExecutionIds against a flattened id→node map. Positive and negative control fixtures verify both directions. Subagent follow-ups on the same commit: - Extract isMissingCandidateActive into graphTraversalUtil so the three async-resolution call sites (model/media pipelines + realtime verification) share a single predicate. - Hoist the flattened id→node map to enrichWithEmbeddedMetadata scope so isModelReferencedByActiveNode doesn't rebuild it per unmatched entry (was O(W*N) allocation). - Rename hasActiveAncestorsInFlattened → isAncestorPathActiveInFlattened for naming symmetry with the live-graph variant. - Document the intentional asymmetry: a candidate dropped by the pipeline-start ancestor filter is not resurrected if the user un-bypasses mid-verification; the realtime mode-change path owns re-surfacing errors after un-bypass. - Unit tests for isMissingCandidateActive (4 cases) plus a positive-control fixture so the Bug 2 regression test cannot pass trivially if a future refactor drops all workflow-level entries.
There was a problem hiding this comment.
-
src/composables/graph/useErrorClearingHooks.ts:237-251
The realtime async verification path still has a stale-result hole across workflow switches.isCandidateStillActive()only re-resolvescandidate.nodeIdagainst the currentapp.rootGraphwhen verification finishes. Unlike the initial-load pipeline, this path does not use an abort controller tied to workflow changes, so a late result from workflow A can be accepted into workflow B whenever execution IDs collide. Root-level IDs like1collide frequently, so this can still resurface missing-model/media errors from a different workflow. -
src/platform/missingModel/missingModelScan.ts:272-344
ThegraphData.modelsfallback is still matched by filename only.isModelReferencedByActiveNode()ignores directory when deciding whether to keep an unmatched workflow-level model alive, so a same-name/different-directory setup can retain the wrong workflow-level missing-model entry. Because these fallback candidates have nonodeId, the later ancestor filter insrc/scripts/app.tscannot remove the false positive.
dante01yoon
left a comment
There was a problem hiding this comment.
Looked through the latest commits — responses to my earlier questions make sense, and the overall approach is solid. One minor item:
missing_models_in_bypassed_subgraph.json — interior CheckpointLoaderSimple node uses "size": [315, 98] and the outer KSampler uses "size": [270, 262], both below the [400, 200] minimum for browser_tests/assets/ fixtures. CodeRabbit already flagged the same issue on missing_models_distinct.json and it was fixed — this fixture should be consistent.
Everything else LGTM — the rootGraphAtScan capture pattern for workflow-switch staleness, the isMissingCandidateActive unified helper, and the ancestor-aware guards all look clean and well-tested.
…0] minimum Matches the browser_tests asset convention (prevents navigateIntoSubgraph timeout issues). Same treatment already applied to missing_models_distinct.json.
Follow-up to #10856. Four correctness issues and their regression tests. ## Bugs fixed ### 1. ErrorOverlay model count reflected node selection `useErrorGroups` exposed `filteredMissingModelGroups` under the public name `missingModelGroups`. `ErrorOverlay.vue` read that alias to compute its model count label, so selecting a node shrank the overlay total. The overlay must always show the whole workflow's errors. Exposed both shapes explicitly: `missingModelGroups` / `missingMediaGroups` (unfiltered totals) and `filteredMissingModelGroups` / `filteredMissingMediaGroups` (selection-scoped). `TabErrors.vue` destructures the filtered variant with an alias. Before https://github.com/user-attachments/assets/eb848c5f-d092-4a4f-b86f-d22bb4408003 After https://github.com/user-attachments/assets/75e67819-c9f2-45ec-9241-74023eca6120 ### 2. Bypass → un-bypass dropped url/hash metadata Realtime `scanNodeModelCandidates` only reads widget values, so un-bypass produced a fresh candidate without the url that `enrichWithEmbeddedMetadata` had previously attached from `graphData.models`. `MissingModelRow`'s download/copy-url buttons disappeared after a bypass/un-bypass cycle. Added `enrichCandidateFromNodeProperties` that copies `url`/`hash`/`directory` from the node's own `properties.models` — which persists across mode toggles — into each scanned candidate. Applied to every call site of the per-node scan. A later fix in the same branch also enforces directory agreement to prevent a same-name / different-directory collision from stamping the wrong metadata. Before https://github.com/user-attachments/assets/39039d83-4d55-41a9-9d01-dec40843741b After https://github.com/user-attachments/assets/047a603b-fb52-4320-886d-dfeed457d833 ### 3. Initial full scan surfaced interior errors of a muted/bypassed subgraph container `scanAllModelCandidates`, `scanAllMediaCandidates`, and the JSON-based missing-node scan only check each node's own mode. Interior nodes whose parent container was bypassed passed the filter. Added `isAncestorPathActive(rootGraph, executionId)` to `graphTraversalUtil` and post-filter the three pipelines in `app.ts` after the live rootGraph is configured. The filter uses the execution-ID path (`"65:63"` → check node 65's mode) so it handles both live-scan-produced and JSON-enrichment-produced candidates. Before https://github.com/user-attachments/assets/3032d46b-81cd-420e-ab8e-f58392267602 After https://github.com/user-attachments/assets/02a01931-951d-4a48-986c-06424044fbf8 ### 4. Bypassed subgraph entry re-surfaced interior errors `useGraphNodeManager` replays `graph.onNodeAdded` for each existing interior node when the Vue node manager initializes on subgraph entry. That chain reached `scanSingleNodeErrors` via `installErrorClearingHooks`' `onNodeAdded` override. Each interior node's own mode was active, so the caller guards passed and the scan re-introduced the error that the initial pipeline had correctly suppressed. Added an ancestor-activity gate at the top of `scanSingleNodeErrors`, the single entry point shared by paste, un-bypass, subgraph entry, and subgraph container activation. A later commit also hardens this guard against detached nodes (null execution ID → skip) and applies the same ancestor check to `isCandidateStillActive` in the realtime verification callback. Before https://github.com/user-attachments/assets/fe44862d-f1d6-41ed-982d-614a7e83d441 After https://github.com/user-attachments/assets/497a76ce-3caa-479f-9024-4cd0f7bd20a4 ## Tests - 6 unit tests for `isAncestorPathActive` (root, active, immediate-bypass, deep-nested mute, unresolvable ancestor, null rootGraph) - 4 unit tests for `enrichCandidateFromNodeProperties` (enrichment, no-overwrite, name mismatch, directory mismatch) - 1 unit test for `scanSingleNodeErrors` ancestor guard (subgraph entry replaying onNodeAdded) - 2 unit tests for `useErrorGroups` dual export + ErrorOverlay contract - 4 E2E tests: - ErrorOverlay model count stays constant when a node is selected (new fixture `missing_models_distinct.json`) - Bypass/un-bypass cycle preserves Copy URL button (uses `missing_models_from_node_properties`) - Loading a workflow with bypassed subgraph suppresses interior missing model error (new fixture `missing_models_in_bypassed_subgraph.json`) - Entering a bypassed subgraph does not resurface interior missing model error (shares the above fixture) `pnpm typecheck`, `pnpm lint`, 206 related unit tests passing. ## Follow-up Several items raised by code review are deferred as pre-existing tech debt or scope-avoided refactors. Tracked via comments on #11215 and #11216. --- Follows up on #10856.
Follow-up to #10856. Four correctness issues and their regression tests. ## Bugs fixed ### 1. ErrorOverlay model count reflected node selection `useErrorGroups` exposed `filteredMissingModelGroups` under the public name `missingModelGroups`. `ErrorOverlay.vue` read that alias to compute its model count label, so selecting a node shrank the overlay total. The overlay must always show the whole workflow's errors. Exposed both shapes explicitly: `missingModelGroups` / `missingMediaGroups` (unfiltered totals) and `filteredMissingModelGroups` / `filteredMissingMediaGroups` (selection-scoped). `TabErrors.vue` destructures the filtered variant with an alias. Before https://github.com/user-attachments/assets/eb848c5f-d092-4a4f-b86f-d22bb4408003 After https://github.com/user-attachments/assets/75e67819-c9f2-45ec-9241-74023eca6120 ### 2. Bypass → un-bypass dropped url/hash metadata Realtime `scanNodeModelCandidates` only reads widget values, so un-bypass produced a fresh candidate without the url that `enrichWithEmbeddedMetadata` had previously attached from `graphData.models`. `MissingModelRow`'s download/copy-url buttons disappeared after a bypass/un-bypass cycle. Added `enrichCandidateFromNodeProperties` that copies `url`/`hash`/`directory` from the node's own `properties.models` — which persists across mode toggles — into each scanned candidate. Applied to every call site of the per-node scan. A later fix in the same branch also enforces directory agreement to prevent a same-name / different-directory collision from stamping the wrong metadata. Before https://github.com/user-attachments/assets/39039d83-4d55-41a9-9d01-dec40843741b After https://github.com/user-attachments/assets/047a603b-fb52-4320-886d-dfeed457d833 ### 3. Initial full scan surfaced interior errors of a muted/bypassed subgraph container `scanAllModelCandidates`, `scanAllMediaCandidates`, and the JSON-based missing-node scan only check each node's own mode. Interior nodes whose parent container was bypassed passed the filter. Added `isAncestorPathActive(rootGraph, executionId)` to `graphTraversalUtil` and post-filter the three pipelines in `app.ts` after the live rootGraph is configured. The filter uses the execution-ID path (`"65:63"` → check node 65's mode) so it handles both live-scan-produced and JSON-enrichment-produced candidates. Before https://github.com/user-attachments/assets/3032d46b-81cd-420e-ab8e-f58392267602 After https://github.com/user-attachments/assets/02a01931-951d-4a48-986c-06424044fbf8 ### 4. Bypassed subgraph entry re-surfaced interior errors `useGraphNodeManager` replays `graph.onNodeAdded` for each existing interior node when the Vue node manager initializes on subgraph entry. That chain reached `scanSingleNodeErrors` via `installErrorClearingHooks`' `onNodeAdded` override. Each interior node's own mode was active, so the caller guards passed and the scan re-introduced the error that the initial pipeline had correctly suppressed. Added an ancestor-activity gate at the top of `scanSingleNodeErrors`, the single entry point shared by paste, un-bypass, subgraph entry, and subgraph container activation. A later commit also hardens this guard against detached nodes (null execution ID → skip) and applies the same ancestor check to `isCandidateStillActive` in the realtime verification callback. Before https://github.com/user-attachments/assets/fe44862d-f1d6-41ed-982d-614a7e83d441 After https://github.com/user-attachments/assets/497a76ce-3caa-479f-9024-4cd0f7bd20a4 ## Tests - 6 unit tests for `isAncestorPathActive` (root, active, immediate-bypass, deep-nested mute, unresolvable ancestor, null rootGraph) - 4 unit tests for `enrichCandidateFromNodeProperties` (enrichment, no-overwrite, name mismatch, directory mismatch) - 1 unit test for `scanSingleNodeErrors` ancestor guard (subgraph entry replaying onNodeAdded) - 2 unit tests for `useErrorGroups` dual export + ErrorOverlay contract - 4 E2E tests: - ErrorOverlay model count stays constant when a node is selected (new fixture `missing_models_distinct.json`) - Bypass/un-bypass cycle preserves Copy URL button (uses `missing_models_from_node_properties`) - Loading a workflow with bypassed subgraph suppresses interior missing model error (new fixture `missing_models_in_bypassed_subgraph.json`) - Entering a bypassed subgraph does not resurface interior missing model error (shares the above fixture) `pnpm typecheck`, `pnpm lint`, 206 related unit tests passing. ## Follow-up Several items raised by code review are deferred as pre-existing tech debt or scope-avoided refactors. Tracked via comments on #11215 and #11216. --- Follows up on #10856.
…10856 (#11466) Backport of #11233 to `core/1.43` Automatically created by backport workflow. ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-11466-backport-core-1-43-fix-missing-asset-correctness-follow-ups-from-10856-3486d73d365081bd94f4e9b8822a8320) by [Unito](https://www.unito.io) Co-authored-by: jaeone94 <89377375+jaeone94@users.noreply.github.com>
…10856 (#11467) Backport of #11233 to `cloud/1.43` Automatically created by backport workflow. ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-11467-backport-cloud-1-43-fix-missing-asset-correctness-follow-ups-from-10856-3486d73d3650811985a7ffef55c0810d) by [Unito](https://www.unito.io) Co-authored-by: jaeone94 <89377375+jaeone94@users.noreply.github.com>
Follow-up to #10856. Four correctness issues and their regression tests. ## Bugs fixed ### 1. ErrorOverlay model count reflected node selection `useErrorGroups` exposed `filteredMissingModelGroups` under the public name `missingModelGroups`. `ErrorOverlay.vue` read that alias to compute its model count label, so selecting a node shrank the overlay total. The overlay must always show the whole workflow's errors. Exposed both shapes explicitly: `missingModelGroups` / `missingMediaGroups` (unfiltered totals) and `filteredMissingModelGroups` / `filteredMissingMediaGroups` (selection-scoped). `TabErrors.vue` destructures the filtered variant with an alias. Before https://github.com/user-attachments/assets/eb848c5f-d092-4a4f-b86f-d22bb4408003 After https://github.com/user-attachments/assets/75e67819-c9f2-45ec-9241-74023eca6120 ### 2. Bypass → un-bypass dropped url/hash metadata Realtime `scanNodeModelCandidates` only reads widget values, so un-bypass produced a fresh candidate without the url that `enrichWithEmbeddedMetadata` had previously attached from `graphData.models`. `MissingModelRow`'s download/copy-url buttons disappeared after a bypass/un-bypass cycle. Added `enrichCandidateFromNodeProperties` that copies `url`/`hash`/`directory` from the node's own `properties.models` — which persists across mode toggles — into each scanned candidate. Applied to every call site of the per-node scan. A later fix in the same branch also enforces directory agreement to prevent a same-name / different-directory collision from stamping the wrong metadata. Before https://github.com/user-attachments/assets/39039d83-4d55-41a9-9d01-dec40843741b After https://github.com/user-attachments/assets/047a603b-fb52-4320-886d-dfeed457d833 ### 3. Initial full scan surfaced interior errors of a muted/bypassed subgraph container `scanAllModelCandidates`, `scanAllMediaCandidates`, and the JSON-based missing-node scan only check each node's own mode. Interior nodes whose parent container was bypassed passed the filter. Added `isAncestorPathActive(rootGraph, executionId)` to `graphTraversalUtil` and post-filter the three pipelines in `app.ts` after the live rootGraph is configured. The filter uses the execution-ID path (`"65:63"` → check node 65's mode) so it handles both live-scan-produced and JSON-enrichment-produced candidates. Before https://github.com/user-attachments/assets/3032d46b-81cd-420e-ab8e-f58392267602 After https://github.com/user-attachments/assets/02a01931-951d-4a48-986c-06424044fbf8 ### 4. Bypassed subgraph entry re-surfaced interior errors `useGraphNodeManager` replays `graph.onNodeAdded` for each existing interior node when the Vue node manager initializes on subgraph entry. That chain reached `scanSingleNodeErrors` via `installErrorClearingHooks`' `onNodeAdded` override. Each interior node's own mode was active, so the caller guards passed and the scan re-introduced the error that the initial pipeline had correctly suppressed. Added an ancestor-activity gate at the top of `scanSingleNodeErrors`, the single entry point shared by paste, un-bypass, subgraph entry, and subgraph container activation. A later commit also hardens this guard against detached nodes (null execution ID → skip) and applies the same ancestor check to `isCandidateStillActive` in the realtime verification callback. Before https://github.com/user-attachments/assets/fe44862d-f1d6-41ed-982d-614a7e83d441 After https://github.com/user-attachments/assets/497a76ce-3caa-479f-9024-4cd0f7bd20a4 ## Tests - 6 unit tests for `isAncestorPathActive` (root, active, immediate-bypass, deep-nested mute, unresolvable ancestor, null rootGraph) - 4 unit tests for `enrichCandidateFromNodeProperties` (enrichment, no-overwrite, name mismatch, directory mismatch) - 1 unit test for `scanSingleNodeErrors` ancestor guard (subgraph entry replaying onNodeAdded) - 2 unit tests for `useErrorGroups` dual export + ErrorOverlay contract - 4 E2E tests: - ErrorOverlay model count stays constant when a node is selected (new fixture `missing_models_distinct.json`) - Bypass/un-bypass cycle preserves Copy URL button (uses `missing_models_from_node_properties`) - Loading a workflow with bypassed subgraph suppresses interior missing model error (new fixture `missing_models_in_bypassed_subgraph.json`) - Entering a bypassed subgraph does not resurface interior missing model error (shares the above fixture) `pnpm typecheck`, `pnpm lint`, 206 related unit tests passing. ## Follow-up Several items raised by code review are deferred as pre-existing tech debt or scope-avoided refactors. Tracked via comments on #11215 and #11216. --- Follows up on #10856.
Follow-up to #10856. Four correctness issues and their regression tests. ## Bugs fixed ### 1. ErrorOverlay model count reflected node selection `useErrorGroups` exposed `filteredMissingModelGroups` under the public name `missingModelGroups`. `ErrorOverlay.vue` read that alias to compute its model count label, so selecting a node shrank the overlay total. The overlay must always show the whole workflow's errors. Exposed both shapes explicitly: `missingModelGroups` / `missingMediaGroups` (unfiltered totals) and `filteredMissingModelGroups` / `filteredMissingMediaGroups` (selection-scoped). `TabErrors.vue` destructures the filtered variant with an alias. Before https://github.com/user-attachments/assets/eb848c5f-d092-4a4f-b86f-d22bb4408003 After https://github.com/user-attachments/assets/75e67819-c9f2-45ec-9241-74023eca6120 ### 2. Bypass → un-bypass dropped url/hash metadata Realtime `scanNodeModelCandidates` only reads widget values, so un-bypass produced a fresh candidate without the url that `enrichWithEmbeddedMetadata` had previously attached from `graphData.models`. `MissingModelRow`'s download/copy-url buttons disappeared after a bypass/un-bypass cycle. Added `enrichCandidateFromNodeProperties` that copies `url`/`hash`/`directory` from the node's own `properties.models` — which persists across mode toggles — into each scanned candidate. Applied to every call site of the per-node scan. A later fix in the same branch also enforces directory agreement to prevent a same-name / different-directory collision from stamping the wrong metadata. Before https://github.com/user-attachments/assets/39039d83-4d55-41a9-9d01-dec40843741b After https://github.com/user-attachments/assets/047a603b-fb52-4320-886d-dfeed457d833 ### 3. Initial full scan surfaced interior errors of a muted/bypassed subgraph container `scanAllModelCandidates`, `scanAllMediaCandidates`, and the JSON-based missing-node scan only check each node's own mode. Interior nodes whose parent container was bypassed passed the filter. Added `isAncestorPathActive(rootGraph, executionId)` to `graphTraversalUtil` and post-filter the three pipelines in `app.ts` after the live rootGraph is configured. The filter uses the execution-ID path (`"65:63"` → check node 65's mode) so it handles both live-scan-produced and JSON-enrichment-produced candidates. Before https://github.com/user-attachments/assets/3032d46b-81cd-420e-ab8e-f58392267602 After https://github.com/user-attachments/assets/02a01931-951d-4a48-986c-06424044fbf8 ### 4. Bypassed subgraph entry re-surfaced interior errors `useGraphNodeManager` replays `graph.onNodeAdded` for each existing interior node when the Vue node manager initializes on subgraph entry. That chain reached `scanSingleNodeErrors` via `installErrorClearingHooks`' `onNodeAdded` override. Each interior node's own mode was active, so the caller guards passed and the scan re-introduced the error that the initial pipeline had correctly suppressed. Added an ancestor-activity gate at the top of `scanSingleNodeErrors`, the single entry point shared by paste, un-bypass, subgraph entry, and subgraph container activation. A later commit also hardens this guard against detached nodes (null execution ID → skip) and applies the same ancestor check to `isCandidateStillActive` in the realtime verification callback. Before https://github.com/user-attachments/assets/fe44862d-f1d6-41ed-982d-614a7e83d441 After https://github.com/user-attachments/assets/497a76ce-3caa-479f-9024-4cd0f7bd20a4 ## Tests - 6 unit tests for `isAncestorPathActive` (root, active, immediate-bypass, deep-nested mute, unresolvable ancestor, null rootGraph) - 4 unit tests for `enrichCandidateFromNodeProperties` (enrichment, no-overwrite, name mismatch, directory mismatch) - 1 unit test for `scanSingleNodeErrors` ancestor guard (subgraph entry replaying onNodeAdded) - 2 unit tests for `useErrorGroups` dual export + ErrorOverlay contract - 4 E2E tests: - ErrorOverlay model count stays constant when a node is selected (new fixture `missing_models_distinct.json`) - Bypass/un-bypass cycle preserves Copy URL button (uses `missing_models_from_node_properties`) - Loading a workflow with bypassed subgraph suppresses interior missing model error (new fixture `missing_models_in_bypassed_subgraph.json`) - Entering a bypassed subgraph does not resurface interior missing model error (shares the above fixture) `pnpm typecheck`, `pnpm lint`, 206 related unit tests passing. ## Follow-up Several items raised by code review are deferred as pre-existing tech debt or scope-avoided refactors. Tracked via comments on #11215 and #11216. --- Follows up on #10856.
Follow-up to #10856. Four correctness issues and their regression tests. ## Bugs fixed ### 1. ErrorOverlay model count reflected node selection `useErrorGroups` exposed `filteredMissingModelGroups` under the public name `missingModelGroups`. `ErrorOverlay.vue` read that alias to compute its model count label, so selecting a node shrank the overlay total. The overlay must always show the whole workflow's errors. Exposed both shapes explicitly: `missingModelGroups` / `missingMediaGroups` (unfiltered totals) and `filteredMissingModelGroups` / `filteredMissingMediaGroups` (selection-scoped). `TabErrors.vue` destructures the filtered variant with an alias. Before https://github.com/user-attachments/assets/eb848c5f-d092-4a4f-b86f-d22bb4408003 After https://github.com/user-attachments/assets/75e67819-c9f2-45ec-9241-74023eca6120 ### 2. Bypass → un-bypass dropped url/hash metadata Realtime `scanNodeModelCandidates` only reads widget values, so un-bypass produced a fresh candidate without the url that `enrichWithEmbeddedMetadata` had previously attached from `graphData.models`. `MissingModelRow`'s download/copy-url buttons disappeared after a bypass/un-bypass cycle. Added `enrichCandidateFromNodeProperties` that copies `url`/`hash`/`directory` from the node's own `properties.models` — which persists across mode toggles — into each scanned candidate. Applied to every call site of the per-node scan. A later fix in the same branch also enforces directory agreement to prevent a same-name / different-directory collision from stamping the wrong metadata. Before https://github.com/user-attachments/assets/39039d83-4d55-41a9-9d01-dec40843741b After https://github.com/user-attachments/assets/047a603b-fb52-4320-886d-dfeed457d833 ### 3. Initial full scan surfaced interior errors of a muted/bypassed subgraph container `scanAllModelCandidates`, `scanAllMediaCandidates`, and the JSON-based missing-node scan only check each node's own mode. Interior nodes whose parent container was bypassed passed the filter. Added `isAncestorPathActive(rootGraph, executionId)` to `graphTraversalUtil` and post-filter the three pipelines in `app.ts` after the live rootGraph is configured. The filter uses the execution-ID path (`"65:63"` → check node 65's mode) so it handles both live-scan-produced and JSON-enrichment-produced candidates. Before https://github.com/user-attachments/assets/3032d46b-81cd-420e-ab8e-f58392267602 After https://github.com/user-attachments/assets/02a01931-951d-4a48-986c-06424044fbf8 ### 4. Bypassed subgraph entry re-surfaced interior errors `useGraphNodeManager` replays `graph.onNodeAdded` for each existing interior node when the Vue node manager initializes on subgraph entry. That chain reached `scanSingleNodeErrors` via `installErrorClearingHooks`' `onNodeAdded` override. Each interior node's own mode was active, so the caller guards passed and the scan re-introduced the error that the initial pipeline had correctly suppressed. Added an ancestor-activity gate at the top of `scanSingleNodeErrors`, the single entry point shared by paste, un-bypass, subgraph entry, and subgraph container activation. A later commit also hardens this guard against detached nodes (null execution ID → skip) and applies the same ancestor check to `isCandidateStillActive` in the realtime verification callback. Before https://github.com/user-attachments/assets/fe44862d-f1d6-41ed-982d-614a7e83d441 After https://github.com/user-attachments/assets/497a76ce-3caa-479f-9024-4cd0f7bd20a4 ## Tests - 6 unit tests for `isAncestorPathActive` (root, active, immediate-bypass, deep-nested mute, unresolvable ancestor, null rootGraph) - 4 unit tests for `enrichCandidateFromNodeProperties` (enrichment, no-overwrite, name mismatch, directory mismatch) - 1 unit test for `scanSingleNodeErrors` ancestor guard (subgraph entry replaying onNodeAdded) - 2 unit tests for `useErrorGroups` dual export + ErrorOverlay contract - 4 E2E tests: - ErrorOverlay model count stays constant when a node is selected (new fixture `missing_models_distinct.json`) - Bypass/un-bypass cycle preserves Copy URL button (uses `missing_models_from_node_properties`) - Loading a workflow with bypassed subgraph suppresses interior missing model error (new fixture `missing_models_in_bypassed_subgraph.json`) - Entering a bypassed subgraph does not resurface interior missing model error (shares the above fixture) `pnpm typecheck`, `pnpm lint`, 206 related unit tests passing. ## Follow-up Several items raised by code review are deferred as pre-existing tech debt or scope-avoided refactors. Tracked via comments on #11215 and #11216. --- Follows up on #10856.
Follow-up to #10856. Four correctness issues and their regression tests.
Bugs fixed
1. ErrorOverlay model count reflected node selection
useErrorGroupsexposedfilteredMissingModelGroupsunder the public namemissingModelGroups.ErrorOverlay.vueread that alias to compute its model count label, so selecting a node shrank the overlay total. The overlay must always show the whole workflow's errors.Exposed both shapes explicitly:
missingModelGroups/missingMediaGroups(unfiltered totals) andfilteredMissingModelGroups/filteredMissingMediaGroups(selection-scoped).TabErrors.vuedestructures the filtered variant with an alias.Before
2026-04-14.162152.mp4
After
2026-04-14.162229.mp4
2. Bypass → un-bypass dropped url/hash metadata
Realtime
scanNodeModelCandidatesonly reads widget values, so un-bypass produced a fresh candidate without the url thatenrichWithEmbeddedMetadatahad previously attached fromgraphData.models.MissingModelRow's download/copy-url buttons disappeared after a bypass/un-bypass cycle.Added
enrichCandidateFromNodePropertiesthat copiesurl/hash/directoryfrom the node's ownproperties.models— which persists across mode toggles — into each scanned candidate. Applied to every call site of the per-node scan. A later fix in the same branch also enforces directory agreement to prevent a same-name / different-directory collision from stamping the wrong metadata.Before
2026-04-14.162350.mp4
After
2026-04-14.162435.mp4
3. Initial full scan surfaced interior errors of a muted/bypassed subgraph container
scanAllModelCandidates,scanAllMediaCandidates, and the JSON-based missing-node scan only check each node's own mode. Interior nodes whose parent container was bypassed passed the filter.Added
isAncestorPathActive(rootGraph, executionId)tographTraversalUtiland post-filter the three pipelines inapp.tsafter the live rootGraph is configured. The filter uses the execution-ID path ("65:63"→ check node 65's mode) so it handles both live-scan-produced and JSON-enrichment-produced candidates.Before
2026-04-14.162555.mp4
After
2026-04-14.162651.mp4
4. Bypassed subgraph entry re-surfaced interior errors
useGraphNodeManagerreplaysgraph.onNodeAddedfor each existing interior node when the Vue node manager initializes on subgraph entry. That chain reachedscanSingleNodeErrorsviainstallErrorClearingHooks'onNodeAddedoverride. Each interior node's own mode was active, so the caller guards passed and the scan re-introduced the error that the initial pipeline had correctly suppressed.Added an ancestor-activity gate at the top of
scanSingleNodeErrors, the single entry point shared by paste, un-bypass, subgraph entry, and subgraph container activation. A later commit also hardens this guard against detached nodes (null execution ID → skip) and applies the same ancestor check toisCandidateStillActivein the realtime verification callback.Before
2026-04-14.162755.mp4
After
2026-04-14.162910.mp4
Tests
isAncestorPathActive(root, active, immediate-bypass, deep-nested mute, unresolvable ancestor, null rootGraph)enrichCandidateFromNodeProperties(enrichment, no-overwrite, name mismatch, directory mismatch)scanSingleNodeErrorsancestor guard (subgraph entry replaying onNodeAdded)useErrorGroupsdual export + ErrorOverlay contractmissing_models_distinct.json)missing_models_from_node_properties)missing_models_in_bypassed_subgraph.json)pnpm typecheck,pnpm lint, 206 related unit tests passing.Follow-up
Several items raised by code review are deferred as pre-existing tech debt or scope-avoided refactors. Tracked via comments on #11215 and #11216.
Follows up on #10856.