@camstack/vision 0.1.0

package/dist/chunk-J3IUBPRE.mjs

@@ -0,0 +1,187 @@
+// src/shared/node-engine.ts
+import * as path from "path";
+var BACKEND_TO_PROVIDER = {
+  cpu: "cpu",
+  coreml: "coreml",
+  cuda: "cuda",
+  tensorrt: "tensorrt",
+  dml: "dml"
+};
+var BACKEND_TO_DEVICE = {
+  cpu: "cpu",
+  coreml: "gpu-mps",
+  cuda: "gpu-cuda",
+  tensorrt: "tensorrt"
+};
+var NodeInferenceEngine = class {
+  constructor(modelPath, backend) {
+    this.modelPath = modelPath;
+    this.backend = backend;
+    this.device = BACKEND_TO_DEVICE[backend] ?? "cpu";
+  }
+  runtime = "onnx";
+  device;
+  session = null;
+  async initialize() {
+    const ort = await import("onnxruntime-node");
+    const provider = BACKEND_TO_PROVIDER[this.backend] ?? "cpu";
+    const absModelPath = path.isAbsolute(this.modelPath) ? this.modelPath : path.resolve(process.cwd(), this.modelPath);
+    const sessionOptions = {
+      executionProviders: [provider]
+    };
+    this.session = await ort.InferenceSession.create(absModelPath, sessionOptions);
+  }
+  async run(input, inputShape) {
+    if (!this.session) {
+      throw new Error("NodeInferenceEngine: not initialized \u2014 call initialize() first");
+    }
+    const ort = await import("onnxruntime-node");
+    const sess = this.session;
+    const inputName = sess.inputNames[0];
+    const tensor = new ort.Tensor("float32", input, [...inputShape]);
+    const feeds = { [inputName]: tensor };
+    const results = await sess.run(feeds);
+    const outputName = sess.outputNames[0];
+    const outputTensor = results[outputName];
+    return outputTensor.data;
+  }
+  async runMultiOutput(input, inputShape) {
+    if (!this.session) {
+      throw new Error("NodeInferenceEngine: not initialized \u2014 call initialize() first");
+    }
+    const ort = await import("onnxruntime-node");
+    const sess = this.session;
+    const inputName = sess.inputNames[0];
+    const tensor = new ort.Tensor("float32", input, [...inputShape]);
+    const feeds = { [inputName]: tensor };
+    const results = await sess.run(feeds);
+    const out = {};
+    for (const name of sess.outputNames) {
+      out[name] = results[name].data;
+    }
+    return out;
+  }
+  async dispose() {
+    this.session = null;
+  }
+};
+
+// src/shared/engine-resolver.ts
+import * as fs from "fs";
+import * as path2 from "path";
+var AUTO_BACKEND_PRIORITY = ["coreml", "cuda", "tensorrt", "cpu"];
+var BACKEND_TO_FORMAT = {
+  cpu: "onnx",
+  coreml: "coreml",
+  cuda: "onnx",
+  tensorrt: "onnx"
+};
+var RUNTIME_TO_FORMAT = {
+  onnx: "onnx",
+  coreml: "coreml",
+  openvino: "openvino",
+  tflite: "tflite",
+  pytorch: "pt"
+};
+function modelFilePath(modelsDir, modelEntry, format) {
+  const formatEntry = modelEntry.formats[format];
+  if (!formatEntry) {
+    throw new Error(`Model ${modelEntry.id} has no ${format} format`);
+  }
+  const urlParts = formatEntry.url.split("/");
+  const filename = urlParts[urlParts.length - 1] ?? `${modelEntry.id}.${format}`;
+  return path2.join(modelsDir, filename);
+}
+function modelExists(filePath) {
+  try {
+    return fs.existsSync(filePath);
+  } catch {
+    return false;
+  }
+}
+async function resolveEngine(options) {
+  const { runtime, backend, modelEntry, modelsDir, downloadModel } = options;
+  let selectedFormat;
+  let selectedBackend;
+  if (runtime === "auto") {
+    const available = await probeOnnxBackends();
+    let chosen = null;
+    for (const b of AUTO_BACKEND_PRIORITY) {
+      if (!available.includes(b)) continue;
+      const fmt = BACKEND_TO_FORMAT[b];
+      if (!fmt) continue;
+      if (!modelEntry.formats[fmt]) continue;
+      chosen = { backend: b, format: fmt };
+      break;
+    }
+    if (!chosen) {
+      throw new Error(
+        `resolveEngine: no compatible backend found for model ${modelEntry.id}. Available backends: ${available.join(", ")}`
+      );
+    }
+    selectedFormat = chosen.format;
+    selectedBackend = chosen.backend;
+  } else {
+    const fmt = RUNTIME_TO_FORMAT[runtime];
+    if (!fmt) {
+      throw new Error(`resolveEngine: unsupported runtime "${runtime}"`);
+    }
+    if (!modelEntry.formats[fmt]) {
+      throw new Error(
+        `resolveEngine: model ${modelEntry.id} has no ${fmt} format for runtime ${runtime}`
+      );
+    }
+    selectedFormat = fmt;
+    selectedBackend = runtime === "onnx" ? backend || "cpu" : runtime;
+  }
+  let modelPath = modelFilePath(modelsDir, modelEntry, selectedFormat);
+  if (!modelExists(modelPath)) {
+    if (downloadModel) {
+      const formatEntry = modelEntry.formats[selectedFormat];
+      modelPath = await downloadModel(formatEntry.url, modelsDir);
+    } else {
+      throw new Error(
+        `resolveEngine: model file not found at ${modelPath} and no downloadModel function provided`
+      );
+    }
+  }
+  if (selectedFormat === "onnx" || selectedFormat === "coreml") {
+    const engine = new NodeInferenceEngine(modelPath, selectedBackend);
+    await engine.initialize();
+    return { engine, format: selectedFormat, modelPath };
+  }
+  const fallbackPath = modelFilePath(modelsDir, modelEntry, "onnx");
+  if (modelEntry.formats["onnx"] && modelExists(fallbackPath)) {
+    const engine = new NodeInferenceEngine(fallbackPath, "cpu");
+    await engine.initialize();
+    return { engine, format: "onnx", modelPath: fallbackPath };
+  }
+  throw new Error(
+    `resolveEngine: format ${selectedFormat} is not yet supported by NodeInferenceEngine and no ONNX fallback is available`
+  );
+}
+async function probeOnnxBackends() {
+  const available = ["cpu"];
+  try {
+    const ort = await import("onnxruntime-node");
+    const providers = ort.env?.webgl?.disabled !== void 0 ? ort.InferenceSession?.getAvailableProviders?.() ?? [] : [];
+    for (const p of providers) {
+      const normalized = p.toLowerCase().replace("executionprovider", "");
+      if (normalized === "coreml") available.push("coreml");
+      else if (normalized === "cuda") available.push("cuda");
+      else if (normalized === "tensorrt") available.push("tensorrt");
+    }
+  } catch {
+  }
+  if (process.platform === "darwin" && !available.includes("coreml")) {
+    available.push("coreml");
+  }
+  return [...new Set(available)];
+}
+
+export {
+  NodeInferenceEngine,
+  resolveEngine,
+  probeOnnxBackends
+};
+//# sourceMappingURL=chunk-J3IUBPRE.mjs.map

package/dist/chunk-J3IUBPRE.mjs.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/shared/node-engine.ts","../src/shared/engine-resolver.ts"],"sourcesContent":["import type { IInferenceEngine, DetectionRuntime, DetectionDevice } from '@camstack/types'\nimport * as path from 'node:path'\n\nconst BACKEND_TO_PROVIDER: Readonly<Record<string, string>> = {\n  cpu: 'cpu',\n  coreml: 'coreml',\n  cuda: 'cuda',\n  tensorrt: 'tensorrt',\n  dml: 'dml',\n} as const\n\nconst BACKEND_TO_DEVICE: Readonly<Record<string, DetectionDevice>> = {\n  cpu: 'cpu',\n  coreml: 'gpu-mps',\n  cuda: 'gpu-cuda',\n  tensorrt: 'tensorrt',\n} as const\n\nexport class NodeInferenceEngine implements IInferenceEngine {\n  readonly runtime: DetectionRuntime = 'onnx'\n  readonly device: DetectionDevice\n  private session: unknown = null\n\n  constructor(\n    private readonly modelPath: string,\n    private readonly backend: string,\n  ) {\n    this.device = (BACKEND_TO_DEVICE[backend] ?? 'cpu') as DetectionDevice\n  }\n\n  async initialize(): Promise<void> {\n    const ort = await import('onnxruntime-node')\n    const provider = BACKEND_TO_PROVIDER[this.backend] ?? 'cpu'\n\n    // Resolve absolute path\n    const absModelPath = path.isAbsolute(this.modelPath)\n      ? this.modelPath\n      : path.resolve(process.cwd(), this.modelPath)\n\n    const sessionOptions: Record<string, unknown> = {\n      executionProviders: [provider],\n    }\n\n    // eslint-disable-next-line @typescript-eslint/no-explicit-any\n    this.session = await (ort as any).InferenceSession.create(absModelPath, sessionOptions)\n  }\n\n  async run(input: Float32Array, inputShape: readonly number[]): Promise<Float32Array> {\n    if (!this.session) {\n      throw new Error('NodeInferenceEngine: not initialized — call initialize() first')\n    }\n\n    const ort = await import('onnxruntime-node')\n    // eslint-disable-next-line @typescript-eslint/no-explicit-any\n    const sess = this.session as any\n\n    // Get the first input name\n    const inputName: string = sess.inputNames[0]\n    // eslint-disable-next-line @typescript-eslint/no-explicit-any\n    const tensor = new (ort as any).Tensor('float32', input, [...inputShape])\n    const feeds: Record<string, unknown> = { [inputName]: tensor }\n\n    const results = await sess.run(feeds)\n    const outputName: string = sess.outputNames[0]\n    const outputTensor = results[outputName]\n\n    return outputTensor.data as Float32Array\n  }\n\n  async runMultiOutput(\n    input: Float32Array,\n    inputShape: readonly number[],\n  ): Promise<Record<string, Float32Array>> {\n    if (!this.session) {\n      throw new Error('NodeInferenceEngine: not initialized — call initialize() first')\n    }\n\n    const ort = await import('onnxruntime-node')\n    // eslint-disable-next-line @typescript-eslint/no-explicit-any\n    const sess = this.session as any\n\n    const inputName: string = sess.inputNames[0]\n    // eslint-disable-next-line @typescript-eslint/no-explicit-any\n    const tensor = new (ort as any).Tensor('float32', input, [...inputShape])\n    const feeds: Record<string, unknown> = { [inputName]: tensor }\n\n    const results = await sess.run(feeds)\n\n    const out: Record<string, Float32Array> = {}\n    for (const name of sess.outputNames as string[]) {\n      out[name] = results[name].data as Float32Array\n    }\n    return out\n  }\n\n  async dispose(): Promise<void> {\n    // onnxruntime-node sessions don't have explicit close in all versions\n    // but we clear the reference\n    this.session = null\n  }\n}\n","import type {\n  IInferenceEngine,\n  DetectionRuntime,\n  ModelCatalogEntry,\n  ModelFormat,\n} from '@camstack/types'\nimport * as fs from 'node:fs'\nimport * as path from 'node:path'\nimport { NodeInferenceEngine } from './node-engine.js'\n\nexport interface EngineResolverOptions {\n  readonly runtime: DetectionRuntime | 'auto'\n  readonly backend: string\n  readonly modelEntry: ModelCatalogEntry\n  readonly modelsDir: string\n  readonly pythonPath?: string\n  readonly downloadModel?: (url: string, destDir: string) => Promise<string>\n}\n\nexport interface ResolvedEngine {\n  readonly engine: IInferenceEngine\n  readonly format: ModelFormat\n  readonly modelPath: string\n}\n\n/** Priority order for auto-selection of ONNX backends */\nconst AUTO_BACKEND_PRIORITY = ['coreml', 'cuda', 'tensorrt', 'cpu'] as const\n\n/** Map backend names to the model format they require */\nconst BACKEND_TO_FORMAT: Readonly<Record<string, ModelFormat>> = {\n  cpu: 'onnx',\n  coreml: 'coreml',\n  cuda: 'onnx',\n  tensorrt: 'onnx',\n} as const\n\n/** Map DetectionRuntime to ModelFormat */\nconst RUNTIME_TO_FORMAT: Readonly<Partial<Record<DetectionRuntime, ModelFormat>>> = {\n  onnx: 'onnx',\n  coreml: 'coreml',\n  openvino: 'openvino',\n  tflite: 'tflite',\n  pytorch: 'pt',\n} as const\n\nfunction modelFilePath(modelsDir: string, modelEntry: ModelCatalogEntry, format: ModelFormat): string {\n  const formatEntry = modelEntry.formats[format]\n  if (!formatEntry) {\n    throw new Error(`Model ${modelEntry.id} has no ${format} format`)\n  }\n  // Derive filename from URL\n  const urlParts = formatEntry.url.split('/')\n  const filename = urlParts[urlParts.length - 1] ?? `${modelEntry.id}.${format}`\n  return path.join(modelsDir, filename)\n}\n\nfunction modelExists(filePath: string): boolean {\n  try {\n    return fs.existsSync(filePath)\n  } catch {\n    return false\n  }\n}\n\nexport async function resolveEngine(options: EngineResolverOptions): Promise<ResolvedEngine> {\n  const { runtime, backend, modelEntry, modelsDir, downloadModel } = options\n\n  let selectedFormat: ModelFormat\n  let selectedBackend: string\n\n  if (runtime === 'auto') {\n    // Probe available ONNX backends and pick best\n    const available = await probeOnnxBackends()\n\n    // Pick first priority backend that has a corresponding model format available\n    let chosen: { backend: string; format: ModelFormat } | null = null\n\n    for (const b of AUTO_BACKEND_PRIORITY) {\n      if (!available.includes(b)) continue\n      const fmt = BACKEND_TO_FORMAT[b]\n      if (!fmt) continue\n      if (!modelEntry.formats[fmt]) continue\n      chosen = { backend: b, format: fmt }\n      break\n    }\n\n    if (!chosen) {\n      throw new Error(\n        `resolveEngine: no compatible backend found for model ${modelEntry.id}. Available backends: ${available.join(', ')}`,\n      )\n    }\n\n    selectedFormat = chosen.format\n    selectedBackend = chosen.backend\n  } else {\n    // Explicit runtime requested\n    const fmt = RUNTIME_TO_FORMAT[runtime]\n    if (!fmt) {\n      throw new Error(`resolveEngine: unsupported runtime \"${runtime}\"`)\n    }\n    if (!modelEntry.formats[fmt]) {\n      throw new Error(\n        `resolveEngine: model ${modelEntry.id} has no ${fmt} format for runtime ${runtime}`,\n      )\n    }\n    selectedFormat = fmt\n    // For onnx runtime, use the provided backend; otherwise use the runtime name\n    selectedBackend = runtime === 'onnx' ? (backend || 'cpu') : runtime\n  }\n\n  // Resolve model path\n  let modelPath = modelFilePath(modelsDir, modelEntry, selectedFormat)\n\n  if (!modelExists(modelPath)) {\n    if (downloadModel) {\n      const formatEntry = modelEntry.formats[selectedFormat]!\n      modelPath = await downloadModel(formatEntry.url, modelsDir)\n    } else {\n      throw new Error(\n        `resolveEngine: model file not found at ${modelPath} and no downloadModel function provided`,\n      )\n    }\n  }\n\n  // Only ONNX runtime is handled by NodeInferenceEngine currently\n  if (selectedFormat === 'onnx' || selectedFormat === 'coreml') {\n    const engine = new NodeInferenceEngine(modelPath, selectedBackend)\n    await engine.initialize()\n    return { engine, format: selectedFormat, modelPath }\n  }\n\n  // For other formats (openvino, tflite, pt), fall back to ONNX cpu with a warning\n  // until those engines are implemented\n  const fallbackPath = modelFilePath(modelsDir, modelEntry, 'onnx')\n  if (modelEntry.formats['onnx'] && modelExists(fallbackPath)) {\n    const engine = new NodeInferenceEngine(fallbackPath, 'cpu')\n    await engine.initialize()\n    return { engine, format: 'onnx', modelPath: fallbackPath }\n  }\n\n  throw new Error(\n    `resolveEngine: format ${selectedFormat} is not yet supported by NodeInferenceEngine and no ONNX fallback is available`,\n  )\n}\n\n/** Probe which ONNX execution providers are available on this system */\nexport async function probeOnnxBackends(): Promise<string[]> {\n  const available: string[] = ['cpu'] // CPU is always available\n\n  try {\n    const ort = await import('onnxruntime-node')\n    // eslint-disable-next-line @typescript-eslint/no-explicit-any\n    const providers: string[] = (ort as any).env?.webgl?.disabled !== undefined\n      // eslint-disable-next-line @typescript-eslint/no-explicit-any\n      ? ((ort as any).InferenceSession?.getAvailableProviders?.() ?? [])\n      : []\n\n    for (const p of providers) {\n      const normalized = p.toLowerCase().replace('executionprovider', '')\n      if (normalized === 'coreml') available.push('coreml')\n      else if (normalized === 'cuda') available.push('cuda')\n      else if (normalized === 'tensorrt') available.push('tensorrt')\n    }\n  } catch {\n    // onnxruntime-node may not be installed; CPU only\n  }\n\n  // Platform-specific hints when getAvailableProviders isn't exposed\n  if (process.platform === 'darwin' && !available.includes('coreml')) {\n    available.push('coreml')\n  }\n\n  return [...new Set(available)]\n}\n"],"mappings":";AACA,YAAY,UAAU;AAEtB,IAAM,sBAAwD;AAAA,EAC5D,KAAK;AAAA,EACL,QAAQ;AAAA,EACR,MAAM;AAAA,EACN,UAAU;AAAA,EACV,KAAK;AACP;AAEA,IAAM,oBAA+D;AAAA,EACnE,KAAK;AAAA,EACL,QAAQ;AAAA,EACR,MAAM;AAAA,EACN,UAAU;AACZ;AAEO,IAAM,sBAAN,MAAsD;AAAA,EAK3D,YACmB,WACA,SACjB;AAFiB;AACA;AAEjB,SAAK,SAAU,kBAAkB,OAAO,KAAK;AAAA,EAC/C;AAAA,EATS,UAA4B;AAAA,EAC5B;AAAA,EACD,UAAmB;AAAA,EAS3B,MAAM,aAA4B;AAChC,UAAM,MAAM,MAAM,OAAO,kBAAkB;AAC3C,UAAM,WAAW,oBAAoB,KAAK,OAAO,KAAK;AAGtD,UAAM,eAAoB,gBAAW,KAAK,SAAS,IAC/C,KAAK,YACA,aAAQ,QAAQ,IAAI,GAAG,KAAK,SAAS;AAE9C,UAAM,iBAA0C;AAAA,MAC9C,oBAAoB,CAAC,QAAQ;AAAA,IAC/B;AAGA,SAAK,UAAU,MAAO,IAAY,iBAAiB,OAAO,cAAc,cAAc;AAAA,EACxF;AAAA,EAEA,MAAM,IAAI,OAAqB,YAAsD;AACnF,QAAI,CAAC,KAAK,SAAS;AACjB,YAAM,IAAI,MAAM,qEAAgE;AAAA,IAClF;AAEA,UAAM,MAAM,MAAM,OAAO,kBAAkB;AAE3C,UAAM,OAAO,KAAK;AAGlB,UAAM,YAAoB,KAAK,WAAW,CAAC;AAE3C,UAAM,SAAS,IAAK,IAAY,OAAO,WAAW,OAAO,CAAC,GAAG,UAAU,CAAC;AACxE,UAAM,QAAiC,EAAE,CAAC,SAAS,GAAG,OAAO;AAE7D,UAAM,UAAU,MAAM,KAAK,IAAI,KAAK;AACpC,UAAM,aAAqB,KAAK,YAAY,CAAC;AAC7C,UAAM,eAAe,QAAQ,UAAU;AAEvC,WAAO,aAAa;AAAA,EACtB;AAAA,EAEA,MAAM,eACJ,OACA,YACuC;AACvC,QAAI,CAAC,KAAK,SAAS;AACjB,YAAM,IAAI,MAAM,qEAAgE;AAAA,IAClF;AAEA,UAAM,MAAM,MAAM,OAAO,kBAAkB;AAE3C,UAAM,OAAO,KAAK;AAElB,UAAM,YAAoB,KAAK,WAAW,CAAC;AAE3C,UAAM,SAAS,IAAK,IAAY,OAAO,WAAW,OAAO,CAAC,GAAG,UAAU,CAAC;AACxE,UAAM,QAAiC,EAAE,CAAC,SAAS,GAAG,OAAO;AAE7D,UAAM,UAAU,MAAM,KAAK,IAAI,KAAK;AAEpC,UAAM,MAAoC,CAAC;AAC3C,eAAW,QAAQ,KAAK,aAAyB;AAC/C,UAAI,IAAI,IAAI,QAAQ,IAAI,EAAE;AAAA,IAC5B;AACA,WAAO;AAAA,EACT;AAAA,EAEA,MAAM,UAAyB;AAG7B,SAAK,UAAU;AAAA,EACjB;AACF;;;AC9FA,YAAY,QAAQ;AACpB,YAAYA,WAAU;AAmBtB,IAAM,wBAAwB,CAAC,UAAU,QAAQ,YAAY,KAAK;AAGlE,IAAM,oBAA2D;AAAA,EAC/D,KAAK;AAAA,EACL,QAAQ;AAAA,EACR,MAAM;AAAA,EACN,UAAU;AACZ;AAGA,IAAM,oBAA8E;AAAA,EAClF,MAAM;AAAA,EACN,QAAQ;AAAA,EACR,UAAU;AAAA,EACV,QAAQ;AAAA,EACR,SAAS;AACX;AAEA,SAAS,cAAc,WAAmB,YAA+B,QAA6B;AACpG,QAAM,cAAc,WAAW,QAAQ,MAAM;AAC7C,MAAI,CAAC,aAAa;AAChB,UAAM,IAAI,MAAM,SAAS,WAAW,EAAE,WAAW,MAAM,SAAS;AAAA,EAClE;AAEA,QAAM,WAAW,YAAY,IAAI,MAAM,GAAG;AAC1C,QAAM,WAAW,SAAS,SAAS,SAAS,CAAC,KAAK,GAAG,WAAW,EAAE,IAAI,MAAM;AAC5E,SAAY,WAAK,WAAW,QAAQ;AACtC;AAEA,SAAS,YAAY,UAA2B;AAC9C,MAAI;AACF,WAAU,cAAW,QAAQ;AAAA,EAC/B,QAAQ;AACN,WAAO;AAAA,EACT;AACF;AAEA,eAAsB,cAAc,SAAyD;AAC3F,QAAM,EAAE,SAAS,SAAS,YAAY,WAAW,cAAc,IAAI;AAEnE,MAAI;AACJ,MAAI;AAEJ,MAAI,YAAY,QAAQ;AAEtB,UAAM,YAAY,MAAM,kBAAkB;AAG1C,QAAI,SAA0D;AAE9D,eAAW,KAAK,uBAAuB;AACrC,UAAI,CAAC,UAAU,SAAS,CAAC,EAAG;AAC5B,YAAM,MAAM,kBAAkB,CAAC;AAC/B,UAAI,CAAC,IAAK;AACV,UAAI,CAAC,WAAW,QAAQ,GAAG,EAAG;AAC9B,eAAS,EAAE,SAAS,GAAG,QAAQ,IAAI;AACnC;AAAA,IACF;AAEA,QAAI,CAAC,QAAQ;AACX,YAAM,IAAI;AAAA,QACR,wDAAwD,WAAW,EAAE,yBAAyB,UAAU,KAAK,IAAI,CAAC;AAAA,MACpH;AAAA,IACF;AAEA,qBAAiB,OAAO;AACxB,sBAAkB,OAAO;AAAA,EAC3B,OAAO;AAEL,UAAM,MAAM,kBAAkB,OAAO;AACrC,QAAI,CAAC,KAAK;AACR,YAAM,IAAI,MAAM,uCAAuC,OAAO,GAAG;AAAA,IACnE;AACA,QAAI,CAAC,WAAW,QAAQ,GAAG,GAAG;AAC5B,YAAM,IAAI;AAAA,QACR,wBAAwB,WAAW,EAAE,WAAW,GAAG,uBAAuB,OAAO;AAAA,MACnF;AAAA,IACF;AACA,qBAAiB;AAEjB,sBAAkB,YAAY,SAAU,WAAW,QAAS;AAAA,EAC9D;AAGA,MAAI,YAAY,cAAc,WAAW,YAAY,cAAc;AAEnE,MAAI,CAAC,YAAY,SAAS,GAAG;AAC3B,QAAI,eAAe;AACjB,YAAM,cAAc,WAAW,QAAQ,cAAc;AACrD,kBAAY,MAAM,cAAc,YAAY,KAAK,SAAS;AAAA,IAC5D,OAAO;AACL,YAAM,IAAI;AAAA,QACR,0CAA0C,SAAS;AAAA,MACrD;AAAA,IACF;AAAA,EACF;AAGA,MAAI,mBAAmB,UAAU,mBAAmB,UAAU;AAC5D,UAAM,SAAS,IAAI,oBAAoB,WAAW,eAAe;AACjE,UAAM,OAAO,WAAW;AACxB,WAAO,EAAE,QAAQ,QAAQ,gBAAgB,UAAU;AAAA,EACrD;AAIA,QAAM,eAAe,cAAc,WAAW,YAAY,MAAM;AAChE,MAAI,WAAW,QAAQ,MAAM,KAAK,YAAY,YAAY,GAAG;AAC3D,UAAM,SAAS,IAAI,oBAAoB,cAAc,KAAK;AAC1D,UAAM,OAAO,WAAW;AACxB,WAAO,EAAE,QAAQ,QAAQ,QAAQ,WAAW,aAAa;AAAA,EAC3D;AAEA,QAAM,IAAI;AAAA,IACR,yBAAyB,cAAc;AAAA,EACzC;AACF;AAGA,eAAsB,oBAAuC;AAC3D,QAAM,YAAsB,CAAC,KAAK;AAElC,MAAI;AACF,UAAM,MAAM,MAAM,OAAO,kBAAkB;AAE3C,UAAM,YAAuB,IAAY,KAAK,OAAO,aAAa,SAE5D,IAAY,kBAAkB,wBAAwB,KAAK,CAAC,IAC9D,CAAC;AAEL,eAAW,KAAK,WAAW;AACzB,YAAM,aAAa,EAAE,YAAY,EAAE,QAAQ,qBAAqB,EAAE;AAClE,UAAI,eAAe,SAAU,WAAU,KAAK,QAAQ;AAAA,eAC3C,eAAe,OAAQ,WAAU,KAAK,MAAM;AAAA,eAC5C,eAAe,WAAY,WAAU,KAAK,UAAU;AAAA,IAC/D;AAAA,EACF,QAAQ;AAAA,EAER;AAGA,MAAI,QAAQ,aAAa,YAAY,CAAC,UAAU,SAAS,QAAQ,GAAG;AAClE,cAAU,KAAK,QAAQ;AAAA,EACzB;AAEA,SAAO,CAAC,GAAG,IAAI,IAAI,SAAS,CAAC;AAC/B;","names":["path"]}

package/dist/chunk-KFZDJPYL.mjs

@@ -0,0 +1,190 @@
+import {
+  cropRegion,
+  resizeAndNormalize
+} from "./chunk-6OR5TE7A.mjs";
+import {
+  resolveEngine
+} from "./chunk-J3IUBPRE.mjs";
+
+// src/addons/face-recognition/index.ts
+import { FACE_RECOGNITION_MODELS } from "@camstack/types";
+
+// src/shared/postprocess/arcface.ts
+function l2Normalize(vec) {
+  let sumSq = 0;
+  for (let i = 0; i < vec.length; i++) {
+    sumSq += vec[i] * vec[i];
+  }
+  const norm = Math.sqrt(sumSq);
+  if (norm === 0) return new Float32Array(vec.length);
+  const out = new Float32Array(vec.length);
+  for (let i = 0; i < vec.length; i++) {
+    out[i] = vec[i] / norm;
+  }
+  return out;
+}
+function cosineSimilarity(a, b) {
+  if (a.length !== b.length) throw new Error("Embedding length mismatch");
+  let dot = 0;
+  for (let i = 0; i < a.length; i++) {
+    dot += a[i] * b[i];
+  }
+  return dot;
+}
+
+// src/addons/face-recognition/index.ts
+var IDENTITY_LABEL = { id: "identity", name: "Identity" };
+var IDENTITY_LABELS = [IDENTITY_LABEL];
+var FACE_REC_CLASS_MAP = { mapping: {}, preserveOriginal: true };
+var REQUIRED_STEPS = [
+  { slot: "cropper", outputClasses: ["face"], description: "Requires a face detector" }
+];
+var FaceRecognitionAddon = class {
+  id = "face-recognition";
+  slot = "classifier";
+  inputClasses = ["face"];
+  outputClasses = ["identity:*"];
+  slotPriority = 0;
+  requiredSteps = REQUIRED_STEPS;
+  manifest = {
+    id: "face-recognition",
+    name: "Face Recognition",
+    version: "0.1.0",
+    description: "ArcFace-based face recognition \u2014 produces 512-d identity embeddings",
+    packageName: "@camstack/vision",
+    slot: "classifier",
+    inputClasses: ["face"],
+    outputClasses: ["identity:*"],
+    requiredSteps: REQUIRED_STEPS,
+    supportsCustomModels: false,
+    mayRequirePython: false,
+    defaultConfig: {
+      modelId: "arcface-r100",
+      runtime: "auto",
+      backend: "cpu"
+    }
+  };
+  engine;
+  modelEntry;
+  async initialize(ctx) {
+    const cfg = ctx.addonConfig;
+    const modelId = cfg["modelId"] ?? "arcface-r100";
+    const runtime = cfg["runtime"] ?? "auto";
+    const backend = cfg["backend"] ?? "cpu";
+    const entry = FACE_RECOGNITION_MODELS.find((m) => m.id === modelId);
+    if (!entry) {
+      throw new Error(`FaceRecognitionAddon: unknown modelId "${modelId}"`);
+    }
+    this.modelEntry = entry;
+    const resolved = await resolveEngine({
+      runtime,
+      backend,
+      modelEntry: entry,
+      modelsDir: ctx.locationPaths.models
+    });
+    this.engine = resolved.engine;
+  }
+  async classify(input) {
+    const start = Date.now();
+    const { width: inputW, height: inputH } = this.modelEntry.inputSize;
+    const faceCrop = await cropRegion(input.frame.data, input.roi);
+    const layout = this.modelEntry.inputLayout ?? "nhwc";
+    const normalization = this.modelEntry.inputNormalization ?? "zero-one";
+    const normalized = await resizeAndNormalize(faceCrop, inputW, inputH, normalization, layout);
+    const rawEmbedding = await this.engine.run(normalized, [1, inputH, inputW, 3]);
+    const embedding = l2Normalize(rawEmbedding);
+    return {
+      classifications: [
+        {
+          class: "identity",
+          score: 1,
+          embedding
+        }
+      ],
+      inferenceMs: Date.now() - start,
+      modelId: this.modelEntry.id
+    };
+  }
+  async shutdown() {
+    await this.engine?.dispose();
+  }
+  getConfigSchema() {
+    return {
+      sections: [
+        {
+          id: "model",
+          title: "Model",
+          columns: 1,
+          fields: [
+            {
+              key: "modelId",
+              label: "Model",
+              type: "model-selector",
+              catalog: [...FACE_RECOGNITION_MODELS],
+              allowCustom: false,
+              allowConversion: false,
+              acceptFormats: ["onnx", "coreml", "openvino"],
+              requiredMetadata: ["inputSize", "labels", "outputFormat"],
+              outputFormatHint: "embedding"
+            }
+          ]
+        },
+        {
+          id: "runtime",
+          title: "Runtime",
+          columns: 2,
+          fields: [
+            {
+              key: "runtime",
+              label: "Runtime",
+              type: "select",
+              options: [
+                { value: "auto", label: "Auto (recommended)" },
+                { value: "onnx", label: "ONNX Runtime" },
+                { value: "coreml", label: "CoreML (Apple)" }
+              ]
+            },
+            {
+              key: "backend",
+              label: "Backend",
+              type: "select",
+              dependsOn: { runtime: "onnx" },
+              options: [
+                { value: "cpu", label: "CPU" },
+                { value: "coreml", label: "CoreML" },
+                { value: "cuda", label: "CUDA (NVIDIA)" }
+              ]
+            }
+          ]
+        }
+      ]
+    };
+  }
+  getClassMap() {
+    return FACE_REC_CLASS_MAP;
+  }
+  getModelCatalog() {
+    return [...FACE_RECOGNITION_MODELS];
+  }
+  getAvailableModels() {
+    return [];
+  }
+  getActiveLabels() {
+    return IDENTITY_LABELS;
+  }
+  async probe() {
+    return {
+      available: true,
+      runtime: this.engine?.runtime ?? "onnx",
+      device: this.engine?.device ?? "cpu",
+      capabilities: ["fp32"]
+    };
+  }
+};
+
+export {
+  l2Normalize,
+  cosineSimilarity,
+  FaceRecognitionAddon
+};
+//# sourceMappingURL=chunk-KFZDJPYL.mjs.map

package/dist/chunk-KFZDJPYL.mjs.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/addons/face-recognition/index.ts","../src/shared/postprocess/arcface.ts"],"sourcesContent":["import type {\n  IClassifierProvider,\n  IDetectionAddon,\n  AddonManifest,\n  AddonContext,\n  CropInput,\n  ClassifierOutput,\n  ConfigUISchema,\n  ClassMapDefinition,\n  ProbeResult,\n  ModelCatalogEntry,\n  DetectionModel,\n  LabelDefinition,\n  IInferenceEngine,\n  RequiredStep,\n} from '@camstack/types'\nimport { FACE_RECOGNITION_MODELS } from '@camstack/types'\nimport { cropRegion, resizeAndNormalize } from '../../shared/image-utils.js'\nimport { l2Normalize } from '../../shared/postprocess/arcface.js'\nimport { resolveEngine } from '../../shared/engine-resolver.js'\n\nconst IDENTITY_LABEL: LabelDefinition = { id: 'identity', name: 'Identity' }\nconst IDENTITY_LABELS: readonly LabelDefinition[] = [IDENTITY_LABEL]\nconst FACE_REC_CLASS_MAP: ClassMapDefinition = { mapping: {}, preserveOriginal: true }\n\nconst REQUIRED_STEPS: readonly RequiredStep[] = [\n  { slot: 'cropper', outputClasses: ['face'], description: 'Requires a face detector' },\n]\n\nexport default class FaceRecognitionAddon implements IClassifierProvider, IDetectionAddon {\n  readonly id = 'face-recognition'\n  readonly slot = 'classifier' as const\n  readonly inputClasses = ['face'] as const\n  readonly outputClasses = ['identity:*'] as const\n  readonly slotPriority = 0\n  readonly requiredSteps = REQUIRED_STEPS\n  readonly manifest: AddonManifest = {\n    id: 'face-recognition',\n    name: 'Face Recognition',\n    version: '0.1.0',\n    description: 'ArcFace-based face recognition — produces 512-d identity embeddings',\n    packageName: '@camstack/vision',\n    slot: 'classifier',\n    inputClasses: ['face'],\n    outputClasses: ['identity:*'],\n    requiredSteps: REQUIRED_STEPS,\n    supportsCustomModels: false,\n    mayRequirePython: false,\n    defaultConfig: {\n      modelId: 'arcface-r100',\n      runtime: 'auto',\n      backend: 'cpu',\n    },\n  }\n\n  private engine!: IInferenceEngine\n  private modelEntry!: ModelCatalogEntry\n\n  async initialize(ctx: AddonContext): Promise<void> {\n    const cfg = ctx.addonConfig\n    const modelId = (cfg['modelId'] as string | undefined) ?? 'arcface-r100'\n    const runtime = (cfg['runtime'] as string | undefined) ?? 'auto'\n    const backend = (cfg['backend'] as string | undefined) ?? 'cpu'\n\n    const entry = FACE_RECOGNITION_MODELS.find((m) => m.id === modelId)\n    if (!entry) {\n      throw new Error(`FaceRecognitionAddon: unknown modelId \"${modelId}\"`)\n    }\n    this.modelEntry = entry\n\n    const resolved = await resolveEngine({\n      runtime: runtime as 'auto',\n      backend,\n      modelEntry: entry,\n      modelsDir: ctx.locationPaths.models,\n    })\n    this.engine = resolved.engine\n  }\n\n  async classify(input: CropInput): Promise<ClassifierOutput> {\n    const start = Date.now()\n    const { width: inputW, height: inputH } = this.modelEntry.inputSize\n\n    // Crop the face region\n    const faceCrop = await cropRegion(input.frame.data, input.roi)\n\n    // Resize to 112x112, NHWC layout (ArcFace expects NHWC)\n    const layout = this.modelEntry.inputLayout ?? 'nhwc'\n    const normalization = this.modelEntry.inputNormalization ?? 'zero-one'\n    const normalized = await resizeAndNormalize(faceCrop, inputW, inputH, normalization, layout)\n\n    // Run inference — output is 512-d embedding\n    const rawEmbedding = await this.engine.run(normalized, [1, inputH, inputW, 3])\n\n    // L2 normalize the embedding\n    const embedding = l2Normalize(rawEmbedding)\n\n    return {\n      classifications: [\n        {\n          class: 'identity',\n          score: 1.0,\n          embedding,\n        },\n      ],\n      inferenceMs: Date.now() - start,\n      modelId: this.modelEntry.id,\n    }\n  }\n\n  async shutdown(): Promise<void> {\n    await this.engine?.dispose()\n  }\n\n  getConfigSchema(): ConfigUISchema {\n    return {\n      sections: [\n        {\n          id: 'model',\n          title: 'Model',\n          columns: 1,\n          fields: [\n            {\n              key: 'modelId',\n              label: 'Model',\n              type: 'model-selector',\n              catalog: [...FACE_RECOGNITION_MODELS],\n              allowCustom: false,\n              allowConversion: false,\n              acceptFormats: ['onnx', 'coreml', 'openvino'],\n              requiredMetadata: ['inputSize', 'labels', 'outputFormat'],\n              outputFormatHint: 'embedding',\n            },\n          ],\n        },\n        {\n          id: 'runtime',\n          title: 'Runtime',\n          columns: 2,\n          fields: [\n            {\n              key: 'runtime',\n              label: 'Runtime',\n              type: 'select',\n              options: [\n                { value: 'auto', label: 'Auto (recommended)' },\n                { value: 'onnx', label: 'ONNX Runtime' },\n                { value: 'coreml', label: 'CoreML (Apple)' },\n              ],\n            },\n            {\n              key: 'backend',\n              label: 'Backend',\n              type: 'select',\n              dependsOn: { runtime: 'onnx' },\n              options: [\n                { value: 'cpu', label: 'CPU' },\n                { value: 'coreml', label: 'CoreML' },\n                { value: 'cuda', label: 'CUDA (NVIDIA)' },\n              ],\n            },\n          ],\n        },\n      ],\n    }\n  }\n\n  getClassMap(): ClassMapDefinition {\n    return FACE_REC_CLASS_MAP\n  }\n\n  getModelCatalog(): ModelCatalogEntry[] {\n    return [...FACE_RECOGNITION_MODELS]\n  }\n\n  getAvailableModels(): DetectionModel[] {\n    return []\n  }\n\n  getActiveLabels(): readonly LabelDefinition[] {\n    return IDENTITY_LABELS\n  }\n\n  async probe(): Promise<ProbeResult> {\n    return {\n      available: true,\n      runtime: this.engine?.runtime ?? 'onnx',\n      device: this.engine?.device ?? 'cpu',\n      capabilities: ['fp32'],\n    }\n  }\n}\n","/** L2 normalize a vector in-place, returning a new Float32Array */\nexport function l2Normalize(vec: Float32Array): Float32Array {\n  let sumSq = 0\n  for (let i = 0; i < vec.length; i++) {\n    sumSq += vec[i]! * vec[i]!\n  }\n  const norm = Math.sqrt(sumSq)\n  if (norm === 0) return new Float32Array(vec.length)\n\n  const out = new Float32Array(vec.length)\n  for (let i = 0; i < vec.length; i++) {\n    out[i] = vec[i]! / norm\n  }\n  return out\n}\n\n/** Cosine similarity between two embeddings (assumes they are already L2-normalized) */\nexport function cosineSimilarity(a: Float32Array, b: Float32Array): number {\n  if (a.length !== b.length) throw new Error('Embedding length mismatch')\n  let dot = 0\n  for (let i = 0; i < a.length; i++) {\n    dot += a[i]! * b[i]!\n  }\n  return dot\n}\n"],"mappings":";;;;;;;;;AAgBA,SAAS,+BAA+B;;;ACfjC,SAAS,YAAY,KAAiC;AAC3D,MAAI,QAAQ;AACZ,WAAS,IAAI,GAAG,IAAI,IAAI,QAAQ,KAAK;AACnC,aAAS,IAAI,CAAC,IAAK,IAAI,CAAC;AAAA,EAC1B;AACA,QAAM,OAAO,KAAK,KAAK,KAAK;AAC5B,MAAI,SAAS,EAAG,QAAO,IAAI,aAAa,IAAI,MAAM;AAElD,QAAM,MAAM,IAAI,aAAa,IAAI,MAAM;AACvC,WAAS,IAAI,GAAG,IAAI,IAAI,QAAQ,KAAK;AACnC,QAAI,CAAC,IAAI,IAAI,CAAC,IAAK;AAAA,EACrB;AACA,SAAO;AACT;AAGO,SAAS,iBAAiB,GAAiB,GAAyB;AACzE,MAAI,EAAE,WAAW,EAAE,OAAQ,OAAM,IAAI,MAAM,2BAA2B;AACtE,MAAI,MAAM;AACV,WAAS,IAAI,GAAG,IAAI,EAAE,QAAQ,KAAK;AACjC,WAAO,EAAE,CAAC,IAAK,EAAE,CAAC;AAAA,EACpB;AACA,SAAO;AACT;;;ADHA,IAAM,iBAAkC,EAAE,IAAI,YAAY,MAAM,WAAW;AAC3E,IAAM,kBAA8C,CAAC,cAAc;AACnE,IAAM,qBAAyC,EAAE,SAAS,CAAC,GAAG,kBAAkB,KAAK;AAErF,IAAM,iBAA0C;AAAA,EAC9C,EAAE,MAAM,WAAW,eAAe,CAAC,MAAM,GAAG,aAAa,2BAA2B;AACtF;AAEA,IAAqB,uBAArB,MAA0F;AAAA,EAC/E,KAAK;AAAA,EACL,OAAO;AAAA,EACP,eAAe,CAAC,MAAM;AAAA,EACtB,gBAAgB,CAAC,YAAY;AAAA,EAC7B,eAAe;AAAA,EACf,gBAAgB;AAAA,EAChB,WAA0B;AAAA,IACjC,IAAI;AAAA,IACJ,MAAM;AAAA,IACN,SAAS;AAAA,IACT,aAAa;AAAA,IACb,aAAa;AAAA,IACb,MAAM;AAAA,IACN,cAAc,CAAC,MAAM;AAAA,IACrB,eAAe,CAAC,YAAY;AAAA,IAC5B,eAAe;AAAA,IACf,sBAAsB;AAAA,IACtB,kBAAkB;AAAA,IAClB,eAAe;AAAA,MACb,SAAS;AAAA,MACT,SAAS;AAAA,MACT,SAAS;AAAA,IACX;AAAA,EACF;AAAA,EAEQ;AAAA,EACA;AAAA,EAER,MAAM,WAAW,KAAkC;AACjD,UAAM,MAAM,IAAI;AAChB,UAAM,UAAW,IAAI,SAAS,KAA4B;AAC1D,UAAM,UAAW,IAAI,SAAS,KAA4B;AAC1D,UAAM,UAAW,IAAI,SAAS,KAA4B;AAE1D,UAAM,QAAQ,wBAAwB,KAAK,CAAC,MAAM,EAAE,OAAO,OAAO;AAClE,QAAI,CAAC,OAAO;AACV,YAAM,IAAI,MAAM,0CAA0C,OAAO,GAAG;AAAA,IACtE;AACA,SAAK,aAAa;AAElB,UAAM,WAAW,MAAM,cAAc;AAAA,MACnC;AAAA,MACA;AAAA,MACA,YAAY;AAAA,MACZ,WAAW,IAAI,cAAc;AAAA,IAC/B,CAAC;AACD,SAAK,SAAS,SAAS;AAAA,EACzB;AAAA,EAEA,MAAM,SAAS,OAA6C;AAC1D,UAAM,QAAQ,KAAK,IAAI;AACvB,UAAM,EAAE,OAAO,QAAQ,QAAQ,OAAO,IAAI,KAAK,WAAW;AAG1D,UAAM,WAAW,MAAM,WAAW,MAAM,MAAM,MAAM,MAAM,GAAG;AAG7D,UAAM,SAAS,KAAK,WAAW,eAAe;AAC9C,UAAM,gBAAgB,KAAK,WAAW,sBAAsB;AAC5D,UAAM,aAAa,MAAM,mBAAmB,UAAU,QAAQ,QAAQ,eAAe,MAAM;AAG3F,UAAM,eAAe,MAAM,KAAK,OAAO,IAAI,YAAY,CAAC,GAAG,QAAQ,QAAQ,CAAC,CAAC;AAG7E,UAAM,YAAY,YAAY,YAAY;AAE1C,WAAO;AAAA,MACL,iBAAiB;AAAA,QACf;AAAA,UACE,OAAO;AAAA,UACP,OAAO;AAAA,UACP;AAAA,QACF;AAAA,MACF;AAAA,MACA,aAAa,KAAK,IAAI,IAAI;AAAA,MAC1B,SAAS,KAAK,WAAW;AAAA,IAC3B;AAAA,EACF;AAAA,EAEA,MAAM,WAA0B;AAC9B,UAAM,KAAK,QAAQ,QAAQ;AAAA,EAC7B;AAAA,EAEA,kBAAkC;AAChC,WAAO;AAAA,MACL,UAAU;AAAA,QACR;AAAA,UACE,IAAI;AAAA,UACJ,OAAO;AAAA,UACP,SAAS;AAAA,UACT,QAAQ;AAAA,YACN;AAAA,cACE,KAAK;AAAA,cACL,OAAO;AAAA,cACP,MAAM;AAAA,cACN,SAAS,CAAC,GAAG,uBAAuB;AAAA,cACpC,aAAa;AAAA,cACb,iBAAiB;AAAA,cACjB,eAAe,CAAC,QAAQ,UAAU,UAAU;AAAA,cAC5C,kBAAkB,CAAC,aAAa,UAAU,cAAc;AAAA,cACxD,kBAAkB;AAAA,YACpB;AAAA,UACF;AAAA,QACF;AAAA,QACA;AAAA,UACE,IAAI;AAAA,UACJ,OAAO;AAAA,UACP,SAAS;AAAA,UACT,QAAQ;AAAA,YACN;AAAA,cACE,KAAK;AAAA,cACL,OAAO;AAAA,cACP,MAAM;AAAA,cACN,SAAS;AAAA,gBACP,EAAE,OAAO,QAAQ,OAAO,qBAAqB;AAAA,gBAC7C,EAAE,OAAO,QAAQ,OAAO,eAAe;AAAA,gBACvC,EAAE,OAAO,UAAU,OAAO,iBAAiB;AAAA,cAC7C;AAAA,YACF;AAAA,YACA;AAAA,cACE,KAAK;AAAA,cACL,OAAO;AAAA,cACP,MAAM;AAAA,cACN,WAAW,EAAE,SAAS,OAAO;AAAA,cAC7B,SAAS;AAAA,gBACP,EAAE,OAAO,OAAO,OAAO,MAAM;AAAA,gBAC7B,EAAE,OAAO,UAAU,OAAO,SAAS;AAAA,gBACnC,EAAE,OAAO,QAAQ,OAAO,gBAAgB;AAAA,cAC1C;AAAA,YACF;AAAA,UACF;AAAA,QACF;AAAA,MACF;AAAA,IACF;AAAA,EACF;AAAA,EAEA,cAAkC;AAChC,WAAO;AAAA,EACT;AAAA,EAEA,kBAAuC;AACrC,WAAO,CAAC,GAAG,uBAAuB;AAAA,EACpC;AAAA,EAEA,qBAAuC;AACrC,WAAO,CAAC;AAAA,EACV;AAAA,EAEA,kBAA8C;AAC5C,WAAO;AAAA,EACT;AAAA,EAEA,MAAM,QAA8B;AAClC,WAAO;AAAA,MACL,WAAW;AAAA,MACX,SAAS,KAAK,QAAQ,WAAW;AAAA,MACjC,QAAQ,KAAK,QAAQ,UAAU;AAAA,MAC/B,cAAc,CAAC,MAAM;AAAA,IACvB;AAAA,EACF;AACF;","names":[]}

package/dist/chunk-KUO2BVFY.mjs

@@ -0,0 +1,90 @@
+// src/shared/postprocess/yolo.ts
+function iou(a, b) {
+  const ax1 = a.x;
+  const ay1 = a.y;
+  const ax2 = a.x + a.w;
+  const ay2 = a.y + a.h;
+  const bx1 = b.x;
+  const by1 = b.y;
+  const bx2 = b.x + b.w;
+  const by2 = b.y + b.h;
+  const interX1 = Math.max(ax1, bx1);
+  const interY1 = Math.max(ay1, by1);
+  const interX2 = Math.min(ax2, bx2);
+  const interY2 = Math.min(ay2, by2);
+  const interW = Math.max(0, interX2 - interX1);
+  const interH = Math.max(0, interY2 - interY1);
+  const interArea = interW * interH;
+  if (interArea === 0) return 0;
+  const areaA = a.w * a.h;
+  const areaB = b.w * b.h;
+  const unionArea = areaA + areaB - interArea;
+  return unionArea === 0 ? 0 : interArea / unionArea;
+}
+function nms(boxes, iouThreshold) {
+  const indices = boxes.map((_, i) => i).sort((a, b) => boxes[b].score - boxes[a].score);
+  const kept = [];
+  const suppressed = /* @__PURE__ */ new Set();
+  for (const idx of indices) {
+    if (suppressed.has(idx)) continue;
+    kept.push(idx);
+    for (const other of indices) {
+      if (other === idx || suppressed.has(other)) continue;
+      if (iou(boxes[idx].bbox, boxes[other].bbox) > iouThreshold) {
+        suppressed.add(other);
+      }
+    }
+  }
+  return kept;
+}
+function yoloPostprocess(output, numClasses, numBoxes, options) {
+  const { confidence, iouThreshold, labels, scale, padX, padY, originalWidth, originalHeight } = options;
+  const candidates = [];
+  for (let i = 0; i < numBoxes; i++) {
+    const cx = output[0 * numBoxes + i];
+    const cy = output[1 * numBoxes + i];
+    const w = output[2 * numBoxes + i];
+    const h = output[3 * numBoxes + i];
+    let bestScore = -Infinity;
+    let bestClass = 0;
+    for (let j = 0; j < numClasses; j++) {
+      const score = output[(4 + j) * numBoxes + i];
+      if (score > bestScore) {
+        bestScore = score;
+        bestClass = j;
+      }
+    }
+    if (bestScore < confidence) continue;
+    const bbox = {
+      x: cx - w / 2,
+      y: cy - h / 2,
+      w,
+      h
+    };
+    candidates.push({ bbox, score: bestScore, classIdx: bestClass });
+  }
+  if (candidates.length === 0) return [];
+  const keptIndices = nms(candidates, iouThreshold);
+  return keptIndices.map((idx) => {
+    const { bbox, score, classIdx } = candidates[idx];
+    const label = labels[classIdx] ?? String(classIdx);
+    const x = Math.max(0, Math.min(originalWidth, (bbox.x - padX) / scale));
+    const y = Math.max(0, Math.min(originalHeight, (bbox.y - padY) / scale));
+    const x2 = Math.max(0, Math.min(originalWidth, (bbox.x + bbox.w - padX) / scale));
+    const y2 = Math.max(0, Math.min(originalHeight, (bbox.y + bbox.h - padY) / scale));
+    const finalBbox = { x, y, w: x2 - x, h: y2 - y };
+    return {
+      class: label,
+      originalClass: label,
+      score,
+      bbox: finalBbox
+    };
+  });
+}
+
+export {
+  iou,
+  nms,
+  yoloPostprocess
+};
+//# sourceMappingURL=chunk-KUO2BVFY.mjs.map

package/dist/chunk-KUO2BVFY.mjs.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/shared/postprocess/yolo.ts"],"sourcesContent":["import type { SpatialDetection, BoundingBox } from '@camstack/types'\n\nexport interface YoloPostprocessOptions {\n  readonly confidence: number\n  readonly iouThreshold: number\n  readonly labels: readonly string[]\n  readonly scale: number\n  readonly padX: number\n  readonly padY: number\n  readonly originalWidth: number\n  readonly originalHeight: number\n}\n\n/** Calculate IoU between two bounding boxes */\nexport function iou(a: BoundingBox, b: BoundingBox): number {\n  const ax1 = a.x\n  const ay1 = a.y\n  const ax2 = a.x + a.w\n  const ay2 = a.y + a.h\n\n  const bx1 = b.x\n  const by1 = b.y\n  const bx2 = b.x + b.w\n  const by2 = b.y + b.h\n\n  const interX1 = Math.max(ax1, bx1)\n  const interY1 = Math.max(ay1, by1)\n  const interX2 = Math.min(ax2, bx2)\n  const interY2 = Math.min(ay2, by2)\n\n  const interW = Math.max(0, interX2 - interX1)\n  const interH = Math.max(0, interY2 - interY1)\n  const interArea = interW * interH\n\n  if (interArea === 0) return 0\n\n  const areaA = a.w * a.h\n  const areaB = b.w * b.h\n  const unionArea = areaA + areaB - interArea\n\n  return unionArea === 0 ? 0 : interArea / unionArea\n}\n\n/** Non-maximum suppression — returns indices of kept boxes (sorted by score desc) */\nexport function nms(\n  boxes: ReadonlyArray<{ readonly bbox: BoundingBox; readonly score: number }>,\n  iouThreshold: number,\n): number[] {\n  const indices = boxes\n    .map((_, i) => i)\n    .sort((a, b) => (boxes[b]!.score) - (boxes[a]!.score))\n\n  const kept: number[] = []\n  const suppressed = new Set<number>()\n\n  for (const idx of indices) {\n    if (suppressed.has(idx)) continue\n    kept.push(idx)\n    for (const other of indices) {\n      if (other === idx || suppressed.has(other)) continue\n      if (iou(boxes[idx]!.bbox, boxes[other]!.bbox) > iouThreshold) {\n        suppressed.add(other)\n      }\n    }\n  }\n\n  return kept\n}\n\n/** Full YOLO v8/v9 postprocessing: filter → NMS → scale back to original coords */\nexport function yoloPostprocess(\n  output: Float32Array,\n  numClasses: number,\n  numBoxes: number,\n  options: YoloPostprocessOptions,\n): SpatialDetection[] {\n  const { confidence, iouThreshold, labels, scale, padX, padY, originalWidth, originalHeight } = options\n\n  interface Candidate {\n    readonly bbox: BoundingBox\n    readonly score: number\n    readonly classIdx: number\n  }\n\n  const candidates: Candidate[] = []\n\n  for (let i = 0; i < numBoxes; i++) {\n    // YOLO v8/v9 output layout: [1, 4+numClasses, numBoxes] stored row-major\n    const cx = output[0 * numBoxes + i]!\n    const cy = output[1 * numBoxes + i]!\n    const w = output[2 * numBoxes + i]!\n    const h = output[3 * numBoxes + i]!\n\n    let bestScore = -Infinity\n    let bestClass = 0\n\n    for (let j = 0; j < numClasses; j++) {\n      const score = output[(4 + j) * numBoxes + i]!\n      if (score > bestScore) {\n        bestScore = score\n        bestClass = j\n      }\n    }\n\n    if (bestScore < confidence) continue\n\n    // Convert cx,cy,w,h to x,y,w,h (top-left origin)\n    const bbox: BoundingBox = {\n      x: cx - w / 2,\n      y: cy - h / 2,\n      w,\n      h,\n    }\n\n    candidates.push({ bbox, score: bestScore, classIdx: bestClass })\n  }\n\n  if (candidates.length === 0) return []\n\n  const keptIndices = nms(candidates, iouThreshold)\n\n  return keptIndices.map((idx) => {\n    const { bbox, score, classIdx } = candidates[idx]!\n    const label = labels[classIdx] ?? String(classIdx)\n\n    // Transform from letterbox coords back to original image coords\n    const x = Math.max(0, Math.min(originalWidth, (bbox.x - padX) / scale))\n    const y = Math.max(0, Math.min(originalHeight, (bbox.y - padY) / scale))\n    const x2 = Math.max(0, Math.min(originalWidth, (bbox.x + bbox.w - padX) / scale))\n    const y2 = Math.max(0, Math.min(originalHeight, (bbox.y + bbox.h - padY) / scale))\n\n    const finalBbox: BoundingBox = { x, y, w: x2 - x, h: y2 - y }\n\n    return {\n      class: label,\n      originalClass: label,\n      score,\n      bbox: finalBbox,\n    } satisfies SpatialDetection\n  })\n}\n"],"mappings":";AAcO,SAAS,IAAI,GAAgB,GAAwB;AAC1D,QAAM,MAAM,EAAE;AACd,QAAM,MAAM,EAAE;AACd,QAAM,MAAM,EAAE,IAAI,EAAE;AACpB,QAAM,MAAM,EAAE,IAAI,EAAE;AAEpB,QAAM,MAAM,EAAE;AACd,QAAM,MAAM,EAAE;AACd,QAAM,MAAM,EAAE,IAAI,EAAE;AACpB,QAAM,MAAM,EAAE,IAAI,EAAE;AAEpB,QAAM,UAAU,KAAK,IAAI,KAAK,GAAG;AACjC,QAAM,UAAU,KAAK,IAAI,KAAK,GAAG;AACjC,QAAM,UAAU,KAAK,IAAI,KAAK,GAAG;AACjC,QAAM,UAAU,KAAK,IAAI,KAAK,GAAG;AAEjC,QAAM,SAAS,KAAK,IAAI,GAAG,UAAU,OAAO;AAC5C,QAAM,SAAS,KAAK,IAAI,GAAG,UAAU,OAAO;AAC5C,QAAM,YAAY,SAAS;AAE3B,MAAI,cAAc,EAAG,QAAO;AAE5B,QAAM,QAAQ,EAAE,IAAI,EAAE;AACtB,QAAM,QAAQ,EAAE,IAAI,EAAE;AACtB,QAAM,YAAY,QAAQ,QAAQ;AAElC,SAAO,cAAc,IAAI,IAAI,YAAY;AAC3C;AAGO,SAAS,IACd,OACA,cACU;AACV,QAAM,UAAU,MACb,IAAI,CAAC,GAAG,MAAM,CAAC,EACf,KAAK,CAAC,GAAG,MAAO,MAAM,CAAC,EAAG,QAAU,MAAM,CAAC,EAAG,KAAM;AAEvD,QAAM,OAAiB,CAAC;AACxB,QAAM,aAAa,oBAAI,IAAY;AAEnC,aAAW,OAAO,SAAS;AACzB,QAAI,WAAW,IAAI,GAAG,EAAG;AACzB,SAAK,KAAK,GAAG;AACb,eAAW,SAAS,SAAS;AAC3B,UAAI,UAAU,OAAO,WAAW,IAAI,KAAK,EAAG;AAC5C,UAAI,IAAI,MAAM,GAAG,EAAG,MAAM,MAAM,KAAK,EAAG,IAAI,IAAI,cAAc;AAC5D,mBAAW,IAAI,KAAK;AAAA,MACtB;AAAA,IACF;AAAA,EACF;AAEA,SAAO;AACT;AAGO,SAAS,gBACd,QACA,YACA,UACA,SACoB;AACpB,QAAM,EAAE,YAAY,cAAc,QAAQ,OAAO,MAAM,MAAM,eAAe,eAAe,IAAI;AAQ/F,QAAM,aAA0B,CAAC;AAEjC,WAAS,IAAI,GAAG,IAAI,UAAU,KAAK;AAEjC,UAAM,KAAK,OAAO,IAAI,WAAW,CAAC;AAClC,UAAM,KAAK,OAAO,IAAI,WAAW,CAAC;AAClC,UAAM,IAAI,OAAO,IAAI,WAAW,CAAC;AACjC,UAAM,IAAI,OAAO,IAAI,WAAW,CAAC;AAEjC,QAAI,YAAY;AAChB,QAAI,YAAY;AAEhB,aAAS,IAAI,GAAG,IAAI,YAAY,KAAK;AACnC,YAAM,QAAQ,QAAQ,IAAI,KAAK,WAAW,CAAC;AAC3C,UAAI,QAAQ,WAAW;AACrB,oBAAY;AACZ,oBAAY;AAAA,MACd;AAAA,IACF;AAEA,QAAI,YAAY,WAAY;AAG5B,UAAM,OAAoB;AAAA,MACxB,GAAG,KAAK,IAAI;AAAA,MACZ,GAAG,KAAK,IAAI;AAAA,MACZ;AAAA,MACA;AAAA,IACF;AAEA,eAAW,KAAK,EAAE,MAAM,OAAO,WAAW,UAAU,UAAU,CAAC;AAAA,EACjE;AAEA,MAAI,WAAW,WAAW,EAAG,QAAO,CAAC;AAErC,QAAM,cAAc,IAAI,YAAY,YAAY;AAEhD,SAAO,YAAY,IAAI,CAAC,QAAQ;AAC9B,UAAM,EAAE,MAAM,OAAO,SAAS,IAAI,WAAW,GAAG;AAChD,UAAM,QAAQ,OAAO,QAAQ,KAAK,OAAO,QAAQ;AAGjD,UAAM,IAAI,KAAK,IAAI,GAAG,KAAK,IAAI,gBAAgB,KAAK,IAAI,QAAQ,KAAK,CAAC;AACtE,UAAM,IAAI,KAAK,IAAI,GAAG,KAAK,IAAI,iBAAiB,KAAK,IAAI,QAAQ,KAAK,CAAC;AACvE,UAAM,KAAK,KAAK,IAAI,GAAG,KAAK,IAAI,gBAAgB,KAAK,IAAI,KAAK,IAAI,QAAQ,KAAK,CAAC;AAChF,UAAM,KAAK,KAAK,IAAI,GAAG,KAAK,IAAI,iBAAiB,KAAK,IAAI,KAAK,IAAI,QAAQ,KAAK,CAAC;AAEjF,UAAM,YAAyB,EAAE,GAAG,GAAG,GAAG,KAAK,GAAG,GAAG,KAAK,EAAE;AAE5D,WAAO;AAAA,MACL,OAAO;AAAA,MACP,eAAe;AAAA,MACf;AAAA,MACA,MAAM;AAAA,IACR;AAAA,EACF,CAAC;AACH;","names":[]}