@camstack/addon-post-analysis 1.2.333 → 1.2.335

package/THIRD_PARTY_MODELS.md

@@ -36,7 +36,7 @@ under [Non-commercial models](#non-commercial-models).
 - RF-DETR © Roboflow, Apache-2.0. https://github.com/roboflow/rf-detr
 - YuNet weights Copyright (c) 2020 Shiqi Yu, MIT. Training code: ShiqiYu/libfacedetection.train, BSD-3-Clause. https://github.com/opencv/opencv_zoo/tree/main/models/face_detection_yunet
 - AuraFace v1 by fal.ai, Apache-2.0. https://huggingface.co/fal/AuraFace-v1
-- © Ultralytics, AGPL-3.0 (YOLOv8n base, fine-tuned for plates by an unidentified third party). CamStack holds no training checkpoint; the ONNX file is the only form it has. https://github.com/ultralytics/ultralytics
+- open-lpr models, Copyright the open-lpr contributors, Apache-2.0 (https://huggingface.co/faisalthaheem/open-lpr-models, tag detector-2026.10.1). YOLOX architecture and COCO pretraining by Megvii (https://github.com/Megvii-BaseDetection/YOLOX), Apache-2.0. The upstream NOTICE is reproduced verbatim (`NOTICE-open-lpr.txt`). https://huggingface.co/faisalthaheem/open-lpr-models
 - fast-plate-ocr, MIT License, copyright (c) 2024 ankandrew (full notice: licenses/models/MIT-fast-plate-ocr.txt). https://github.com/ankandrew/fast-plate-ocr
 - EfficientNet-Lite0 via timm (Ross Wightman), Apache-2.0. Open Images V7 annotations © Google LLC, CC BY 4.0; images listed by Google as CC BY 2.0, without warranty. https://github.com/huggingface/pytorch-image-models
 - torchvision MobileNetV3, BSD-3-Clause. Trained on the Vehicle classification dataset by DrBimmer (https://huggingface.co/datasets/DrBimmer/vehicle-classification), CC BY-NC 4.0. https://github.com/pytorch/vision
@@ -62,34 +62,42 @@ Training data under attribution terms:
   cards link here instead of repeating it.
 -->
 
-The folders `objectDetection/yolo26`, `objectDetection/yolov9`, `plateDetection/yolov8-plate` and `segmentation/yolo26-seg` of the model mirror hold Ultralytics weights,
+The folders `objectDetection/yolo26`, `objectDetection/yolov9` and `segmentation/yolo26-seg` of the model mirror hold Ultralytics weights,
 modified by CamStack and distributed under AGPL-3.0. Their Corresponding Source
 is published in the mirror itself, in its **`source/`** folder,
 **https://huggingface.co/camstack/camstack-models/tree/main/source**, together
 with the upstream checkpoints it starts from:
 
-1. **The upstream checkpoints**, published by Ultralytics at
-   https://github.com/ultralytics/assets/releases: `yolo26{n,s,m,l,x}.pt`,
-   `yolov9{t,s,m,c}.pt`, `yolo26{n,s,m}-seg.pt`.
-2. **The export scripts** in `source/`: `build-camstack-models.py` (ONNX
-   with a dynamic batch axis, CoreML FP16, OpenVINO IR, OpenVINO NNCF INT8),
-   `convert_lib.py` (the ONNX to OpenVINO / INT8 / CoreML converter it imports,
-   the same one CamStack's Model Studio runs), `requirements.txt` and
-   `export-models.py`. The build instructions (Python environment,
-   invocation, calibration images for INT8) are in `source/README.md` and in
-   each script's header. These scripts are CamStack's own code and are offered
-   under the MIT licence; the weights they produce keep their upstream licence.
-
-What the published scripts do **not** cover yet:
-
-- **Some hosted sizes and input resolutions.** The export scripts do not list
-  every variant in the mirror (for example the `-320`/`-256` YOLO26 exports,
-  `yolov9m`/`yolov9c`, and the `yolo26*-seg` models). Each one is the same
-  Ultralytics `export()` with a different model id or `imgsz`, but no
-  published script reproduces it as-is.
-- **`yolov8n-plate` has no training checkpoint.** It is a third-party model
-  (trained 2023-12-05 according to its own metadata, origin not identified).
-  The ONNX file is the only form CamStack has.
+1. **The upstream checkpoints.** Ultralytics publishes
+   `yolo26{n,s,m,l,x}.pt`, `yolov9{t,s,m,c}.pt` and `yolo26{n,s,m}-seg.pt` at
+   https://github.com/ultralytics/assets/releases.
+2. **The build script** in `source/`: `build-camstack-models.py`, with
+   `convert_lib.py` (the ONNX to OpenVINO / INT8 converter it calls, the same
+   file CamStack's Model Studio runs) and `requirements.txt`. It holds one
+   recipe per hosted file: the upstream checkpoint, the input size, the
+   Ultralytics `export()` arguments, the OpenVINO precision and the INT8
+   calibration count that file records about itself. Every hosted
+   Ultralytics-derived file is rebuilt by one invocation:
+   `build-camstack-models.py --variant <file stem>` (for example
+   `camstack-yolo26n-320` or `camstack-yolov9m-640`), `--family <folder>`
+   or `--all`; `--list` prints the recipes and `--check-hosted` confirms
+   against the mirror that every hosted file is a recipe output. The INT8
+   builds are NNCF post-training quantisation calibrated on COCO images: 128
+   for YOLO26, 300 for YOLOv9, as each IR records; the documented set is the
+   first N of COCO val2017 in file-name order. The Python
+   environment, the recorded tool versions and the calibration set are in
+   `source/README.md` and the script's header. These scripts are CamStack's
+   own code and are offered under the MIT licence; the weights they produce
+   keep their upstream licence.
+
+Limits of a rebuild:
+
+- **Tool versions.** With the Ultralytics, PyTorch, coremltools and OpenVINO
+  versions each recipe names, a rebuild reproduces the hosted weights. Other
+  versions give the same weights in a possibly different graph.
+- **INT8 calibration.** An IR records how many images calibrated it, not
+  which ones. Calibrating on the same COCO images gives the documented build;
+  a different set gives an equivalent, not bit-identical, INT8 model.
 
 ## Non-commercial models
 
@@ -138,7 +146,7 @@ entry kept resolvable for a saved selection but no longer offered.
 
 | Catalog ids | Upstream | Code / weights | Training data | CamStack changes | Downloaded from |
 | --- | --- | --- | --- | --- | --- |
-| `yolov8n-plate` (default) | [Ultralytics](https://github.com/ultralytics/ultralytics) | AGPL-3.0 / AGPL-3.0 (`AGPL-3.0.txt`) | unidentified plate dataset (third-party model, trained 2023-12-05 per its ONNX metadata): UNKNOWN | converted to OpenVINO IR and CoreML | mirror `plateDetection/yolov8-plate` |
+| `yolox-tiny-plate` (default) | [open-lpr models (faisalthaheem/open-lpr-models), YOLOX-tiny](https://huggingface.co/faisalthaheem/open-lpr-models) | Apache-2.0 / Apache-2.0 (`Apache-2.0.txt`) | Simanno annotated plate corpus plus community-contributed photographs, pseudo-labelled by a teacher model (not published; the upstream NOTICE says the photographs were uploaded without consent obtained): UNKNOWN | input contract baked into the graph (static 1x3x640x640 RGB /255 becomes the upstream BGR 0..255: Mul(255) plus a channel Gather) and the output reshaped to the YOLOv8 layout (cx, cy, w, h, obj*cls); ONNX fp32, OpenVINO IR fp16, CoreML fp16 (scripts/build-yolox-plate-model.py) | mirror `plateDetection/yolox-tiny-plate` |
 
 ### Licence plate reading
 
@@ -156,7 +164,7 @@ entry kept resolvable for a saved selection but no longer offered.
 
 | Catalog ids | Upstream | Code / weights | Training data | CamStack changes | Downloaded from |
 | --- | --- | --- | --- | --- | --- |
-| `bird-classifier` (default) | [Google AIY Vision birds_V1](https://github.com/google-coral/test_data) | — / Apache-2.0 (`Apache-2.0.txt`) | — | converted from TensorFlow with tf2onnx; OpenVINO IR and CoreML derived from it | mirror `animalClassification/bird-classifier` |
+| `bird-classifier` (default) | [Google AIY Vision birds_V1](https://github.com/google-coral/test_data) | — / Apache-2.0 (`Apache-2.0.txt`) | — | converted from TensorFlow with tf2onnx; OpenVINO IR and CoreML derived from it; the Edge TPU build is the Google full-integer quantised TFLite of the same model, compiled with edgetpu_compiler 14 (scripts/build-bird-classifier-edgetpu.py) | mirror `animalClassification/bird-classifier` |
 
 ### Vehicle classifier
 

package/dist/{clip-model-registry-BEh9LdQS.js → clip-model-registry-BxUHFGUg.js}

@@ -11560,6 +11560,90 @@ function resolveClipSearchMinScore(meta, request) {
 	const factor = CLIP_STRICTNESS_FACTOR[request.strictness ?? "loose"];
 	return Math.min(1, Math.round(meta.searchMinScore * factor * 1e3) / 1e3);
 }
+/**
+* The class names a per-class COUNT can ever be keyed on — the macro ids, and
+* only those.
+*
+* `preserveOriginal: false` below is the whole reason this set exists: the
+* executor maps each raw COCO label onto its macro and DROPS everything else,
+* so a tracked detection's `className` is always one of {@link MACRO_LABELS}.
+* Any surface that indexes counts by class name — the ZoneAnalytics occupancy
+* snapshot's `byClass`, and therefore every occupancy rule — can match nothing
+* else. Derived from the catalog, never hand-listed, so a new macro joins it by
+* being declared once.
+*
+* Cost of not having it, 2026-08-12: the notification rule editor offered the
+* full taxonomy for an occupancy class (occupancy is the one condition that
+* does not go through the engine's macro-expanding `expandClassSelector`), two
+* live rules were authored on `car`, and both counters read 0 forever.
+*/
+var DETECTION_MACRO_CLASSES = new Set([
+	{
+		id: "person",
+		name: "Person"
+	},
+	{
+		id: "vehicle",
+		name: "Vehicle"
+	},
+	{
+		id: "animal",
+		name: "Animal"
+	},
+	{
+		id: "package",
+		name: "Package"
+	}
+].map((l) => l.id));
+/** True when `className` is a macro the detector can actually emit (and hence a
+*  class a count can be keyed on) — see {@link DETECTION_MACRO_CLASSES}. */
+function isDetectionMacroClass(className) {
+	return DETECTION_MACRO_CLASSES.has(className.trim().toLowerCase());
+}
+var COCO_TO_MACRO = {
+	mapping: {
+		person: "person",
+		bicycle: "vehicle",
+		car: "vehicle",
+		motorcycle: "vehicle",
+		airplane: "vehicle",
+		bus: "vehicle",
+		train: "vehicle",
+		truck: "vehicle",
+		boat: "vehicle",
+		bird: "animal",
+		cat: "animal",
+		dog: "animal",
+		horse: "animal",
+		sheep: "animal",
+		cow: "animal",
+		elephant: "animal",
+		bear: "animal",
+		zebra: "animal",
+		giraffe: "animal",
+		suitcase: "package",
+		backpack: "package",
+		handbag: "package"
+	},
+	preserveOriginal: false
+};
+/** Steps a root detector may declare it fills. */
+var FILLABLE_STEP_IDS = [
+	"animal-classifier",
+	"vehicle-classifier",
+	"package-detection"
+];
+/** Narrow a step id to a fillable one — type guard, not a cast. */
+function isFillableStepId(value) {
+	return FILLABLE_STEP_IDS.some((id) => id === value);
+}
+/** The step a linked label stands in for, or `undefined` for any other attribution. */
+function linkedStepOf(attributionStepId) {
+	if (attributionStepId === void 0) return void 0;
+	if (!attributionStepId.startsWith("detector-link:")) return void 0;
+	const stepId = attributionStepId.slice(14);
+	return isFillableStepId(stepId) ? stepId : void 0;
+}
 var MODEL_FORMATS = [
 	"onnx",
 	"coreml",
@@ -11692,6 +11776,12 @@ var ModelLicenseSchema = preprocess((value) => typeof value === "string" ? { wei
 	/** True when the weights or their data forbid commercial use. */
 	nonCommercial: boolean().optional()
 }));
+/** One `fillsClassifierEffects` entry — see `ModelCatalogEntry.fillsClassifierEffects`. */
+var FillsClassifierEffectSchema = object({
+	stepId: _enum(FILLABLE_STEP_IDS),
+	labelToEffect: record(string(), string()),
+	uncovered: array(string()).readonly()
+});
 var ModelCatalogEntrySchema = object({
 	id: string(),
 	name: string(),
@@ -11818,6 +11908,16 @@ var ModelCatalogEntrySchema = object({
 	*/
 	classMap: DetectionCatalogClassMapSchema.optional(),
 	/**
+	* The pipeline steps this ROOT detector already does the work of (D715):
+	* its labels name the effect the step exists to produce (`deer` for
+	* `animal-classifier`, `package` for `package-detection`). Derived from the
+	* labelmap by `deriveClassifierFills`, never hand-written; a COCO model
+	* declares none. Whether a declaration is honoured is the step's own
+	* `labelLink` policy (`resolveLabelLink`). `labelToEffect` values are in
+	* `FILLABLE_STEP_VOCABULARIES` (guard `check-label-link-vocabulary`).
+	*/
+	fillsClassifierEffects: array(FillsClassifierEffectSchema).readonly().optional(),
+	/**
 	* The model's CLIP feature-space contract (text encoder, dimension, vector
 	* index, token window, search floor) — see `clip-model.ts`. Present ONLY on a
 	* CLIP image encoder; its absence is what "not a CLIP model" means to the
@@ -21460,73 +21560,6 @@ var notificationOutputCapability = {
 		})
 	}
 };
-/**
-* The class names a per-class COUNT can ever be keyed on — the macro ids, and
-* only those.
-*
-* `preserveOriginal: false` below is the whole reason this set exists: the
-* executor maps each raw COCO label onto its macro and DROPS everything else,
-* so a tracked detection's `className` is always one of {@link MACRO_LABELS}.
-* Any surface that indexes counts by class name — the ZoneAnalytics occupancy
-* snapshot's `byClass`, and therefore every occupancy rule — can match nothing
-* else. Derived from the catalog, never hand-listed, so a new macro joins it by
-* being declared once.
-*
-* Cost of not having it, 2026-08-12: the notification rule editor offered the
-* full taxonomy for an occupancy class (occupancy is the one condition that
-* does not go through the engine's macro-expanding `expandClassSelector`), two
-* live rules were authored on `car`, and both counters read 0 forever.
-*/
-var DETECTION_MACRO_CLASSES = new Set([
-	{
-		id: "person",
-		name: "Person"
-	},
-	{
-		id: "vehicle",
-		name: "Vehicle"
-	},
-	{
-		id: "animal",
-		name: "Animal"
-	},
-	{
-		id: "package",
-		name: "Package"
-	}
-].map((l) => l.id));
-/** True when `className` is a macro the detector can actually emit (and hence a
-*  class a count can be keyed on) — see {@link DETECTION_MACRO_CLASSES}. */
-function isDetectionMacroClass(className) {
-	return DETECTION_MACRO_CLASSES.has(className.trim().toLowerCase());
-}
-var COCO_TO_MACRO = {
-	mapping: {
-		person: "person",
-		bicycle: "vehicle",
-		car: "vehicle",
-		motorcycle: "vehicle",
-		airplane: "vehicle",
-		bus: "vehicle",
-		train: "vehicle",
-		truck: "vehicle",
-		boat: "vehicle",
-		bird: "animal",
-		cat: "animal",
-		dog: "animal",
-		horse: "animal",
-		sheep: "animal",
-		cow: "animal",
-		elephant: "animal",
-		bear: "animal",
-		zebra: "animal",
-		giraffe: "animal",
-		suitcase: "package",
-		backpack: "package",
-		handbag: "package"
-	},
-	preserveOriginal: false
-};
 var AUDIO_MACRO_LABELS = [
 	{
 		id: "speech",
@@ -22331,6 +22364,8 @@ var NcSystemEventKindSchema = _enum([
 	"stream-offline",
 	"node-online",
 	"node-offline",
+	"node-restarting",
+	"node-updating",
 	"node-inference-unavailable",
 	"detection-blind",
 	"addon-crash-loop",
@@ -23814,6 +23849,14 @@ var NC_CONDITION_CATALOG = [
 				value: "node-offline",
 				label: "Node offline"
 			},
+			{
+				value: "node-restarting",
+				label: "Node restarting"
+			},
+			{
+				value: "node-updating",
+				label: "Node updating (restart imminent)"
+			},
 			{
 				value: "node-inference-unavailable",
 				label: "Node has no usable inference device"
@@ -28616,7 +28659,8 @@ var PipelineModelOptionSchema = object({
 	group: ModelVariantGroupSchema.optional(),
 	legacy: boolean().optional(),
 	provider: ModelProviderIdSchema.optional(),
-	defaultConfidence: number().optional()
+	defaultConfidence: number().optional(),
+	fillsClassifierEffects: array(FillsClassifierEffectSchema).readonly().optional()
 });
 var ConfigFieldBridge = custom();
 var PipelineAddonSchemaSchema = object({
@@ -40063,7 +40107,8 @@ var MotionStatusSchema = object({
 	*   writes this slice (not even a boot seed), and neither does the runner's
 	*   phase machine: its phase is "a session is attached" for every on-motion
 	*   session, not motion, and it now has its own `DetectionPhaseTransition`.
-	*   A provider's `isDetected` answers from the hub's slice.
+	*   A provider's `isDetected` and `getStatus` answer from the hub's slice
+	*   (`hubMotionNativeProvider`).
 	* - **Sensors** (Home Assistant binary sensors, Homematic, Matter, Alexa
 	*   imports) — the provider pushes both edges itself, from the upstream
 	*   system's own state change. That is the sensor's own fact.
@@ -60150,7 +60195,7 @@ function systemEventFilterApplies(kind, filter) {
 		case "deviceIds": return kind.startsWith("device-") || kind.startsWith("stream-") || kind === "detection-blind" || kind === "alarm-triggered";
 		case "deviceTypes": return kind.startsWith("device-") || kind === "detection-blind";
 		case "nodeIds": return kind.startsWith("node-") || kind === "addon-crash-loop" || kind === "addon-update-available" || kind === "addon-updated" || kind === "server-update-available" || kind === "wrapper-update-available" || kind === "server-updated";
-		case "packageNames": return kind.endsWith("update-available") || kind === "addon-updated" || kind === "server-updated";
+		case "packageNames": return kind.endsWith("update-available") || kind === "addon-updated" || kind === "node-updating" || kind === "server-updated";
 		case "consumableLowPercent": return NC_CONSUMABLE_SYSTEM_EVENT_KINDS.includes(kind);
 	}
 }
@@ -60195,6 +60240,8 @@ var NC_SYSTEM_EVENT_FAMILY = {
 	"detection-blind": "device-events",
 	"node-online": "system",
 	"node-offline": "system",
+	"node-restarting": "system",
+	"node-updating": "system",
 	"node-inference-unavailable": "system",
 	"addon-crash-loop": "system",
 	"addon-update-available": "system",
@@ -60290,6 +60337,7 @@ var NC_SYSTEM_EVENT_KIND_GROUP_SPECS = [
 		kinds: [
 			"node-offline",
 			"node-online",
+			"node-restarting",
 			"node-inference-unavailable"
 		]
 	},
@@ -60311,6 +60359,7 @@ var NC_SYSTEM_EVENT_KIND_GROUP_SPECS = [
 		label: "Server & container updates",
 		kinds: [
 			"server-update-available",
+			"node-updating",
 			"server-updated",
 			"wrapper-update-available"
 		]
@@ -61217,6 +61266,7 @@ var STATE_SUMMARY_SECTION_VARS = [
 		kinds: [
 			"node-offline",
 			"node-online",
+			"node-restarting",
 			"node-inference-unavailable"
 		]
 	},
@@ -61238,6 +61288,7 @@ var STATE_SUMMARY_SECTION_VARS = [
 		example: "server 1.2.377 → 1.2.378",
 		kinds: [
 			"server-update-available",
+			"node-updating",
 			"server-updated",
 			"wrapper-update-available"
 		]
@@ -61643,6 +61694,8 @@ var NC_TEMPLATE_VARS = [
 		systemEventKinds: [
 			"node-online",
 			"node-offline",
+			"node-restarting",
+			"node-updating",
 			"node-inference-unavailable",
 			"addon-crash-loop",
 			...UPDATES,
@@ -61661,6 +61714,8 @@ var NC_TEMPLATE_VARS = [
 		systemEventKinds: [
 			"node-online",
 			"node-offline",
+			"node-restarting",
+			"node-updating",
 			"node-inference-unavailable",
 			"addon-crash-loop",
 			...UPDATES,
@@ -61714,6 +61769,7 @@ var NC_TEMPLATE_VARS = [
 		systemEventKinds: [
 			...UPDATES,
 			"addon-updated",
+			"node-updating",
 			"server-updated"
 		]
 	},
@@ -61725,7 +61781,11 @@ var NC_TEMPLATE_VARS = [
 		families: ["rule"],
 		fields: [...RULE_TEXT],
 		kinds: ["system"],
-		systemEventKinds: [...PACKAGE_UPDATES, "server-updated"]
+		systemEventKinds: [
+			...PACKAGE_UPDATES,
+			"node-updating",
+			"server-updated"
+		]
 	},
 	{
 		name: "nodes",
@@ -61777,7 +61837,7 @@ var NC_TEMPLATE_VARS = [
 		families: ["rule"],
 		fields: [...RULE_TEXT],
 		kinds: ["system"],
-		systemEventKinds: ["server-updated"]
+		systemEventKinds: ["node-updating", "server-updated"]
 	},
 	{
 		name: "toVersion",
@@ -61787,7 +61847,7 @@ var NC_TEMPLATE_VARS = [
 		families: ["rule"],
 		fields: [...RULE_TEXT],
 		kinds: ["system"],
-		systemEventKinds: ["server-updated"]
+		systemEventKinds: ["node-updating", "server-updated"]
 	},
 	{
 		name: "destinationsOk",
@@ -62056,7 +62116,9 @@ var NC_TEMPLATE_VARS = [
 var NC_TEMPLATE_HELD_BACK_SYSTEM_EVENT_KINDS = [
 	"device-tank-low",
 	"device-tank-full",
-	"device-tank-normal"
+	"device-tank-normal",
+	"node-restarting",
+	"node-updating"
 ];
 /**
 * The catalog as `getTemplateCatalog` SERVES it: every descriptor, with the
@@ -64993,6 +65055,12 @@ Object.defineProperty(exports, "kebabToCamel", {
 		return kebabToCamel;
 	}
 });
+Object.defineProperty(exports, "linkedStepOf", {
+	enumerable: true,
+	get: function() {
+		return linkedStepOf;
+	}
+});
 Object.defineProperty(exports, "literal", {
 	enumerable: true,
 	get: function() {