Repository navigation
Add structured DeepCORO result renderers - #453
Conversation
There was a problem hiding this comment.
Sorry @JacquesDelfrate, you've used your own review budget of 250,000 diff characters for the last 7 days.
You can request another review in 4 days and 19 hours by commenting @sourcery-ai review. Upgrade to get a review now.
Reviewer's GuideThe PR fixes deployed DeepCORO result rendering by explicitly recognizing deployed CLIP and MACE display identities alongside existing canonical names, while preserving exact supported-version and payload validation; registry, renderer, drawer, and contract documentation coverage is updated accordingly. Flow diagram for DeepCORO result renderer selectionflowchart LR
R[Model execution result] --> I{Recognized model identity?}
I -->|DeepCoro_CLIP_generic or DeepCORO-CLIP| C{Version 1.0.0 and valid payload?}
I -->|DeepCORO_MACE or DeepCORO-MACE| M{Version 1.0.0 and valid payload?}
I -->|Other identity| G[Generic renderer]
C -->|Yes| CR[DeepCORO-CLIP renderer]
C -->|No| G
M -->|Yes| MR[DeepCORO-MACE renderer]
M -->|No| G
File-Level Changes
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
There was a problem hiding this comment.
Risk: Low. Approved.
The diff only adds exact display-name aliases (DeepCORO-CLIP, DeepCORO-MACE) next to the existing canonical names in the DeepCORO result renderer matchers. Supported version remains 1.0.0, and the existing payload parsers are unchanged. A name match with an invalid payload still falls through to the generic renderer. Other model adapters, auth, infrastructure, and data models are untouched.
No CODEOWNERS file and no reviewers were already assigned. No additional reviewers requested.
Sent by Cursor Automation: Assign PR reviewers
There was a problem hiding this comment.
Risk: Medium. Prior approval withdrawn.
The earlier Low-risk approval covered only exact display-name aliases for DeepCORO-CLIP and DeepCORO-MACE. Commit e837e880 adds production parsers and result views for DeepCORO-CTO 2.0.0 and DeepCORO-SYNTAX 5.0.0, and registers both on the shared model-result renderer. Users now see J-CTO score, difficulty, selected artery, modified SYNTAX, territory, severity probabilities, and a research threshold for those model identities. Payloads that fail validation still fall through to the generic renderer. There is no auth, infrastructure, or schema change, and there is no CODEOWNERS file.
The approval on 5da51865 was dismissed when this commit landed. This update is not approved.
Reviewers requested: danvincent11 and kabaluyot.
Review focus:
- CTO parsing requires
modelRecommendations.presentable, and the result view never reads it, so a non-presentable payload still shows scores. - Diagnosis and recommendation strings are parsed and not rendered. SYNTAX shows the allowlisted band text, and CTO shows the allowlisted difficulty text.
- Optional
ctoArteryWarningtext is validated and not shown. The LCx caution is a fixed string whenever the selected artery is LCx. - Confirm deployed units: J-CTO in 0–4, SYNTAX and territory in 0–100 points, component and severity probabilities in 0–1.
Sent by Cursor Automation: Assign PR reviewers


Summary:
Validation:
Summary by Sourcery
Support deployed DeepCORO model identities and present validated CTO and SYNTAX results with dedicated, safety-focused renderers.
New Features:
Bug Fixes:
Enhancements:
Documentation:
Tests: