@taodigital/crawler 0.0.0-stage → 1.2.1

package/dist/openapi-generator.js

@@ -0,0 +1,89 @@
+function extractPathParams(urlPattern) {
+  const params = [];
+  const matches = urlPattern.matchAll(/\{(\w+)\}/g);
+  for (const m of matches) {
+    params.push({ name: m[1], in: "path", required: true, schema: { type: "string" } });
+  }
+  return params;
+}
+function methodSummary(method, path) {
+  const resource = path.split("/").filter((s) => !s.startsWith("{") && s.length > 0).pop() ?? "resource";
+  const summaries = {
+    GET: `Get ${resource}`,
+    POST: `Create ${resource}`,
+    PUT: `Update ${resource}`,
+    PATCH: `Partially update ${resource}`,
+    DELETE: `Delete ${resource}`
+  };
+  return summaries[method.toUpperCase()] ?? `${method} ${resource}`;
+}
+function inferTag(path) {
+  const parts = path.split("/").filter((s) => s.length > 0 && !s.startsWith("{"));
+  return parts[0] ?? "default";
+}
+function generateOpenApiSpec(observations, opts = {}) {
+  const { title = "ZeTa Discovered API", baseUrl = "/", version = "1.0.0" } = opts;
+  const spec = {
+    openapi: "3.0.3",
+    info: { title, version, description: "Auto-generated by ZeTa crawler from observed network traffic." },
+    servers: [{ url: baseUrl }],
+    paths: {}
+  };
+  for (const obs of observations) {
+    const path = obs.urlPattern;
+    const method = obs.method.toLowerCase();
+    if (!spec.paths[path]) spec.paths[path] = {};
+    const params = extractPathParams(path);
+    const statusCode = obs.statusCode ?? 200;
+    const responses = {
+      [String(statusCode)]: {
+        description: statusCode < 400 ? "Successful response" : "Error response",
+        content: obs.responseShape ? { "application/json": { schema: shapeToJsonSchema(obs.responseShape) } } : void 0
+      }
+    };
+    const op = {
+      summary: methodSummary(obs.method, path),
+      parameters: params.length > 0 ? params : void 0,
+      responses,
+      tags: [inferTag(path)]
+    };
+    spec.paths[path][method] = op;
+  }
+  return spec;
+}
+function shapeToJsonSchema(shape) {
+  if (Array.isArray(shape)) return { type: "array", items: shape.length > 0 ? shapeToJsonSchema(shape[0]) : {} };
+  if (shape && typeof shape === "object") {
+    const props = {};
+    for (const [k, v] of Object.entries(shape)) props[k] = shapeToJsonSchema(v);
+    return { type: "object", properties: props };
+  }
+  return { type: String(shape) };
+}
+function specToYaml(spec) {
+  function toYaml(val, indent = 0) {
+    const pad = "  ".repeat(indent);
+    if (val === null || val === void 0) return "null";
+    if (typeof val === "string") return val.includes("\n") || val.includes(":") ? `"${val.replace(/"/g, '\\"')}"` : val;
+    if (typeof val === "number" || typeof val === "boolean") return String(val);
+    if (Array.isArray(val)) return val.map((v) => `
+${pad}- ${toYaml(v, indent + 1).trimStart()}`).join("") || "[]";
+    if (typeof val === "object") {
+      const entries = Object.entries(val);
+      if (entries.length === 0) return "{}";
+      return entries.map(([k, v]) => {
+        const vStr = toYaml(v, indent + 1);
+        return typeof v === "object" && v !== null && !Array.isArray(v) ? `
+${pad}${k}:${vStr}` : `
+${pad}${k}: ${vStr.trimStart()}`;
+      }).join("");
+    }
+    return String(val);
+  }
+  return `openapi: "${spec.openapi}"${toYaml(spec, 0)}`;
+}
+export {
+  generateOpenApiSpec,
+  specToYaml
+};
+//# sourceMappingURL=openapi-generator.js.map

package/dist/openapi-generator.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/openapi-generator.ts"],
+  "sourcesContent": ["/**\n * OpenAPI 3.0 spec generator from observed network traffic.\n *\n * Takes the ApiEndpointObservationInput[] collected by the network observer\n * during crawl and generates a minimal but valid OpenAPI 3.0 spec.\n *\n * The spec captures: endpoints, methods, observed response shapes, status codes.\n * It does NOT capture request bodies (we'd need to intercept requests too) \u2014 that's\n * marked as a TODO. But it's enough for generating smoke-test suites.\n */\n\nimport type { ApiEndpointObservationInput } from './network-observer.js';\n\nexport interface OpenApiSpec {\n  openapi: '3.0.3';\n  info: { title: string; version: string; description?: string };\n  servers: Array<{ url: string }>;\n  paths: Record<string, Record<string, OpenApiOperation>>;\n}\n\ninterface OpenApiOperation {\n  summary: string;\n  parameters?: OpenApiParam[];\n  responses: Record<string, { description: string; content?: Record<string, { schema: unknown }> }>;\n  tags?: string[];\n}\n\ninterface OpenApiParam {\n  name: string;\n  in: 'path' | 'query';\n  required: boolean;\n  schema: { type: string };\n}\n\nfunction extractPathParams(urlPattern: string): OpenApiParam[] {\n  const params: OpenApiParam[] = [];\n  const matches = urlPattern.matchAll(/\\{(\\w+)\\}/g);\n  for (const m of matches) {\n    params.push({ name: m[1], in: 'path', required: true, schema: { type: 'string' } });\n  }\n  return params;\n}\n\nfunction methodSummary(method: string, path: string): string {\n  const resource = path.split('/').filter(s => !s.startsWith('{') && s.length > 0).pop() ?? 'resource';\n  const summaries: Record<string, string> = {\n    GET: `Get ${resource}`,\n    POST: `Create ${resource}`,\n    PUT: `Update ${resource}`,\n    PATCH: `Partially update ${resource}`,\n    DELETE: `Delete ${resource}`,\n  };\n  return summaries[method.toUpperCase()] ?? `${method} ${resource}`;\n}\n\nfunction inferTag(path: string): string {\n  const parts = path.split('/').filter(s => s.length > 0 && !s.startsWith('{'));\n  return parts[0] ?? 'default';\n}\n\nexport function generateOpenApiSpec(\n  observations: ApiEndpointObservationInput[],\n  opts: { title?: string; baseUrl?: string; version?: string } = {},\n): OpenApiSpec {\n  const { title = 'ZeTa Discovered API', baseUrl = '/', version = '1.0.0' } = opts;\n\n  const spec: OpenApiSpec = {\n    openapi: '3.0.3',\n    info: { title, version, description: 'Auto-generated by ZeTa crawler from observed network traffic.' },\n    servers: [{ url: baseUrl }],\n    paths: {},\n  };\n\n  for (const obs of observations) {\n    const path = obs.urlPattern;\n    const method = obs.method.toLowerCase();\n\n    if (!spec.paths[path]) spec.paths[path] = {};\n\n    const params = extractPathParams(path);\n    const statusCode = obs.statusCode ?? 200;\n    const responses: OpenApiOperation['responses'] = {\n      [String(statusCode)]: {\n        description: statusCode < 400 ? 'Successful response' : 'Error response',\n        content: obs.responseShape\n          ? { 'application/json': { schema: shapeToJsonSchema(obs.responseShape) } }\n          : undefined,\n      },\n    };\n\n    const op: OpenApiOperation = {\n      summary: methodSummary(obs.method, path),\n      parameters: params.length > 0 ? params : undefined,\n      responses,\n      tags: [inferTag(path)],\n    };\n\n    spec.paths[path][method] = op;\n  }\n\n  return spec;\n}\n\nfunction shapeToJsonSchema(shape: unknown): unknown {\n  if (Array.isArray(shape)) return { type: 'array', items: shape.length > 0 ? shapeToJsonSchema(shape[0]) : {} };\n  if (shape && typeof shape === 'object') {\n    const props: Record<string, unknown> = {};\n    for (const [k, v] of Object.entries(shape as object)) props[k] = shapeToJsonSchema(v);\n    return { type: 'object', properties: props };\n  }\n  return { type: String(shape) };\n}\n\nexport function specToYaml(spec: OpenApiSpec): string {\n  // Minimal YAML serializer \u2014 avoids yaml dep\n  function toYaml(val: unknown, indent = 0): string {\n    const pad = '  '.repeat(indent);\n    if (val === null || val === undefined) return 'null';\n    if (typeof val === 'string') return val.includes('\\n') || val.includes(':') ? `\"${val.replace(/\"/g, '\\\\\"')}\"` : val;\n    if (typeof val === 'number' || typeof val === 'boolean') return String(val);\n    if (Array.isArray(val)) return val.map(v => `\\n${pad}- ${toYaml(v, indent + 1).trimStart()}`).join('') || '[]';\n    if (typeof val === 'object') {\n      const entries = Object.entries(val as object);\n      if (entries.length === 0) return '{}';\n      return entries.map(([k, v]) => {\n        const vStr = toYaml(v, indent + 1);\n        return typeof v === 'object' && v !== null && !Array.isArray(v)\n          ? `\\n${pad}${k}:${vStr}`\n          : `\\n${pad}${k}: ${vStr.trimStart()}`;\n      }).join('');\n    }\n    return String(val);\n  }\n  return `openapi: \"${spec.openapi}\"${toYaml(spec, 0)}`;\n}\n"],
+  "mappings": "AAkCA,SAAS,kBAAkB,YAAoC;AAC7D,QAAM,SAAyB,CAAC;AAChC,QAAM,UAAU,WAAW,SAAS,YAAY;AAChD,aAAW,KAAK,SAAS;AACvB,WAAO,KAAK,EAAE,MAAM,EAAE,CAAC,GAAG,IAAI,QAAQ,UAAU,MAAM,QAAQ,EAAE,MAAM,SAAS,EAAE,CAAC;AAAA,EACpF;AACA,SAAO;AACT;AAEA,SAAS,cAAc,QAAgB,MAAsB;AAC3D,QAAM,WAAW,KAAK,MAAM,GAAG,EAAE,OAAO,OAAK,CAAC,EAAE,WAAW,GAAG,KAAK,EAAE,SAAS,CAAC,EAAE,IAAI,KAAK;AAC1F,QAAM,YAAoC;AAAA,IACxC,KAAK,OAAO,QAAQ;AAAA,IACpB,MAAM,UAAU,QAAQ;AAAA,IACxB,KAAK,UAAU,QAAQ;AAAA,IACvB,OAAO,oBAAoB,QAAQ;AAAA,IACnC,QAAQ,UAAU,QAAQ;AAAA,EAC5B;AACA,SAAO,UAAU,OAAO,YAAY,CAAC,KAAK,GAAG,MAAM,IAAI,QAAQ;AACjE;AAEA,SAAS,SAAS,MAAsB;AACtC,QAAM,QAAQ,KAAK,MAAM,GAAG,EAAE,OAAO,OAAK,EAAE,SAAS,KAAK,CAAC,EAAE,WAAW,GAAG,CAAC;AAC5E,SAAO,MAAM,CAAC,KAAK;AACrB;AAEO,SAAS,oBACd,cACA,OAA+D,CAAC,GACnD;AACb,QAAM,EAAE,QAAQ,uBAAuB,UAAU,KAAK,UAAU,QAAQ,IAAI;AAE5E,QAAM,OAAoB;AAAA,IACxB,SAAS;AAAA,IACT,MAAM,EAAE,OAAO,SAAS,aAAa,gEAAgE;AAAA,IACrG,SAAS,CAAC,EAAE,KAAK,QAAQ,CAAC;AAAA,IAC1B,OAAO,CAAC;AAAA,EACV;AAEA,aAAW,OAAO,cAAc;AAC9B,UAAM,OAAO,IAAI;AACjB,UAAM,SAAS,IAAI,OAAO,YAAY;AAEtC,QAAI,CAAC,KAAK,MAAM,IAAI,EAAG,MAAK,MAAM,IAAI,IAAI,CAAC;AAE3C,UAAM,SAAS,kBAAkB,IAAI;AACrC,UAAM,aAAa,IAAI,cAAc;AACrC,UAAM,YAA2C;AAAA,MAC/C,CAAC,OAAO,UAAU,CAAC,GAAG;AAAA,QACpB,aAAa,aAAa,MAAM,wBAAwB;AAAA,QACxD,SAAS,IAAI,gBACT,EAAE,oBAAoB,EAAE,QAAQ,kBAAkB,IAAI,aAAa,EAAE,EAAE,IACvE;AAAA,MACN;AAAA,IACF;AAEA,UAAM,KAAuB;AAAA,MAC3B,SAAS,cAAc,IAAI,QAAQ,IAAI;AAAA,MACvC,YAAY,OAAO,SAAS,IAAI,SAAS;AAAA,MACzC;AAAA,MACA,MAAM,CAAC,SAAS,IAAI,CAAC;AAAA,IACvB;AAEA,SAAK,MAAM,IAAI,EAAE,MAAM,IAAI;AAAA,EAC7B;AAEA,SAAO;AACT;AAEA,SAAS,kBAAkB,OAAyB;AAClD,MAAI,MAAM,QAAQ,KAAK,EAAG,QAAO,EAAE,MAAM,SAAS,OAAO,MAAM,SAAS,IAAI,kBAAkB,MAAM,CAAC,CAAC,IAAI,CAAC,EAAE;AAC7G,MAAI,SAAS,OAAO,UAAU,UAAU;AACtC,UAAM,QAAiC,CAAC;AACxC,eAAW,CAAC,GAAG,CAAC,KAAK,OAAO,QAAQ,KAAe,EAAG,OAAM,CAAC,IAAI,kBAAkB,CAAC;AACpF,WAAO,EAAE,MAAM,UAAU,YAAY,MAAM;AAAA,EAC7C;AACA,SAAO,EAAE,MAAM,OAAO,KAAK,EAAE;AAC/B;AAEO,SAAS,WAAW,MAA2B;AAEpD,WAAS,OAAO,KAAc,SAAS,GAAW;AAChD,UAAM,MAAM,KAAK,OAAO,MAAM;AAC9B,QAAI,QAAQ,QAAQ,QAAQ,OAAW,QAAO;AAC9C,QAAI,OAAO,QAAQ,SAAU,QAAO,IAAI,SAAS,IAAI,KAAK,IAAI,SAAS,GAAG,IAAI,IAAI,IAAI,QAAQ,MAAM,KAAK,CAAC,MAAM;AAChH,QAAI,OAAO,QAAQ,YAAY,OAAO,QAAQ,UAAW,QAAO,OAAO,GAAG;AAC1E,QAAI,MAAM,QAAQ,GAAG,EAAG,QAAO,IAAI,IAAI,OAAK;AAAA,EAAK,GAAG,KAAK,OAAO,GAAG,SAAS,CAAC,EAAE,UAAU,CAAC,EAAE,EAAE,KAAK,EAAE,KAAK;AAC1G,QAAI,OAAO,QAAQ,UAAU;AAC3B,YAAM,UAAU,OAAO,QAAQ,GAAa;AAC5C,UAAI,QAAQ,WAAW,EAAG,QAAO;AACjC,aAAO,QAAQ,IAAI,CAAC,CAAC,GAAG,CAAC,MAAM;AAC7B,cAAM,OAAO,OAAO,GAAG,SAAS,CAAC;AACjC,eAAO,OAAO,MAAM,YAAY,MAAM,QAAQ,CAAC,MAAM,QAAQ,CAAC,IAC1D;AAAA,EAAK,GAAG,GAAG,CAAC,IAAI,IAAI,KACpB;AAAA,EAAK,GAAG,GAAG,CAAC,KAAK,KAAK,UAAU,CAAC;AAAA,MACvC,CAAC,EAAE,KAAK,EAAE;AAAA,IACZ;AACA,WAAO,OAAO,GAAG;AAAA,EACnB;AACA,SAAO,aAAa,KAAK,OAAO,IAAI,OAAO,MAAM,CAAC,CAAC;AACrD;",
+  "names": []
+}

package/dist/pdf-extractor.d.ts

@@ -0,0 +1,28 @@
+/**
+ * PDF extraction — discovers and extracts text from PDF URLs found during crawl.
+ *
+ * When the HTTP engine or browser engine discovers a link to a .pdf file,
+ * this module fetches the PDF, extracts its text content, and returns it
+ * as a screen-compatible markdown string for test generation.
+ *
+ * Uses pdf-parse (already installed in API package, accessible via dynamic import).
+ */
+export interface PdfScreen {
+    url: string;
+    title: string;
+    markdown: string;
+    pageCount: number;
+    info: Record<string, string>;
+    sizeBytes: number;
+}
+export declare function isPdfUrl(url: string, contentType?: string): boolean;
+export declare function extractPdfFromUrl(url: string, opts?: {
+    userAgent?: string;
+    timeoutMs?: number;
+    maxSizeBytes?: number;
+}): Promise<PdfScreen | null>;
+/**
+ * Filter a list of discovered URLs to find PDF links.
+ * Used by http-engine and browser crawler to decide which links to pass to extractPdfFromUrl.
+ */
+export declare function filterPdfUrls(urls: string[]): string[];

package/dist/pdf-extractor.js

@@ -0,0 +1,81 @@
+const PDF_EXT = /\.pdf(\?.*)?$/i;
+const PDF_MIME = "application/pdf";
+function isPdfUrl(url, contentType) {
+  return PDF_EXT.test(url) || (contentType?.includes(PDF_MIME) ?? false);
+}
+async function extractPdfFromUrl(url, opts = {}) {
+  const { userAgent = "ZeTa-Crawler/1.0", timeoutMs = 3e4, maxSizeBytes = 50 * 1024 * 1024 } = opts;
+  let buffer;
+  let sizeBytes;
+  try {
+    const response = await fetch(url, {
+      headers: { "User-Agent": userAgent, "Accept": "application/pdf,*/*" },
+      signal: AbortSignal.timeout(timeoutMs)
+    });
+    if (!response.ok) return null;
+    const contentType = response.headers.get("content-type") ?? "";
+    if (!contentType.includes("pdf") && !PDF_EXT.test(url)) return null;
+    const arrayBuffer = await response.arrayBuffer();
+    sizeBytes = arrayBuffer.byteLength;
+    if (sizeBytes > maxSizeBytes) {
+      console.warn(`[pdf-extractor] PDF too large (${sizeBytes} bytes), skipping: ${url}`);
+      return null;
+    }
+    buffer = Buffer.from(arrayBuffer);
+  } catch (e) {
+    console.error(`[pdf-extractor] Failed to fetch PDF ${url}: ${e.message}`);
+    return null;
+  }
+  try {
+    const pdfParse = await import("pdf-parse/lib/pdf-parse.js").catch(
+      () => import("pdf-parse")
+    );
+    const parse = pdfParse.default ?? pdfParse;
+    const data = await parse(buffer, { max: 0 });
+    const title = data.info?.Title || extractTitleFromUrl(url);
+    const markdown = pdfTextToMarkdown(data.text ?? "", title);
+    return {
+      url,
+      title,
+      markdown,
+      pageCount: data.numpages ?? 0,
+      info: {
+        author: data.info?.Author ?? "",
+        subject: data.info?.Subject ?? "",
+        keywords: data.info?.Keywords ?? "",
+        creator: data.info?.Creator ?? ""
+      },
+      sizeBytes
+    };
+  } catch (e) {
+    console.error(`[pdf-extractor] Failed to parse PDF ${url}: ${e.message}`);
+    return null;
+  }
+}
+function extractTitleFromUrl(url) {
+  try {
+    const { pathname } = new URL(url);
+    const filename = pathname.split("/").pop() ?? "document";
+    return filename.replace(/\.pdf$/i, "").replace(/[-_]/g, " ").trim();
+  } catch {
+    return "PDF Document";
+  }
+}
+function pdfTextToMarkdown(text, title) {
+  if (!text.trim()) return `# ${title}
+
+*No extractable text content.*`;
+  const cleaned = text.replace(/\f/g, "\n\n---\n\n").replace(/\r\n/g, "\n").replace(/\r/g, "\n").replace(/\n{3,}/g, "\n\n").replace(/[ \t]{2,}/g, " ").trim();
+  return `# ${title}
+
+${cleaned}`;
+}
+function filterPdfUrls(urls) {
+  return urls.filter((u) => PDF_EXT.test(u));
+}
+export {
+  extractPdfFromUrl,
+  filterPdfUrls,
+  isPdfUrl
+};
+//# sourceMappingURL=pdf-extractor.js.map

package/dist/pdf-extractor.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/pdf-extractor.ts"],
+  "sourcesContent": ["/**\n * PDF extraction \u2014 discovers and extracts text from PDF URLs found during crawl.\n *\n * When the HTTP engine or browser engine discovers a link to a .pdf file,\n * this module fetches the PDF, extracts its text content, and returns it\n * as a screen-compatible markdown string for test generation.\n *\n * Uses pdf-parse (already installed in API package, accessible via dynamic import).\n */\n\nconst PDF_EXT = /\\.pdf(\\?.*)?$/i;\nconst PDF_MIME = 'application/pdf';\n\nexport interface PdfScreen {\n  url: string;\n  title: string;\n  markdown: string;\n  pageCount: number;\n  info: Record<string, string>;\n  sizeBytes: number;\n}\n\nexport function isPdfUrl(url: string, contentType?: string): boolean {\n  return PDF_EXT.test(url) || (contentType?.includes(PDF_MIME) ?? false);\n}\n\nexport async function extractPdfFromUrl(url: string, opts: {\n  userAgent?: string;\n  timeoutMs?: number;\n  maxSizeBytes?: number;\n} = {}): Promise<PdfScreen | null> {\n  const { userAgent = 'ZeTa-Crawler/1.0', timeoutMs = 30_000, maxSizeBytes = 50 * 1024 * 1024 } = opts;\n\n  let buffer: Buffer;\n  let sizeBytes: number;\n\n  try {\n    const response = await fetch(url, {\n      headers: { 'User-Agent': userAgent, 'Accept': 'application/pdf,*/*' },\n      signal: AbortSignal.timeout(timeoutMs),\n    });\n    if (!response.ok) return null;\n\n    const contentType = response.headers.get('content-type') ?? '';\n    if (!contentType.includes('pdf') && !PDF_EXT.test(url)) return null;\n\n    const arrayBuffer = await response.arrayBuffer();\n    sizeBytes = arrayBuffer.byteLength;\n    if (sizeBytes > maxSizeBytes) {\n      console.warn(`[pdf-extractor] PDF too large (${sizeBytes} bytes), skipping: ${url}`);\n      return null;\n    }\n    buffer = Buffer.from(arrayBuffer);\n  } catch (e: any) {\n    console.error(`[pdf-extractor] Failed to fetch PDF ${url}: ${e.message}`);\n    return null;\n  }\n\n  try {\n    // pdf-parse is installed in the API package \u2014 use dynamic import\n    // so this file compiles without adding pdf-parse to crawler's package.json\n    // eslint-disable-next-line @typescript-eslint/ban-ts-comment\n    // @ts-ignore \u2013 no types for pdf-parse/lib/pdf-parse.js\n    const pdfParse = await import('pdf-parse/lib/pdf-parse.js').catch(\n      () => import('pdf-parse' as any)\n    ) as any;\n    const parse = pdfParse.default ?? pdfParse;\n    const data = await parse(buffer, { max: 0 }); // max: 0 = all pages\n\n    const title = (data.info?.Title as string) || extractTitleFromUrl(url);\n    const markdown = pdfTextToMarkdown(data.text ?? '', title);\n\n    return {\n      url,\n      title,\n      markdown,\n      pageCount: data.numpages ?? 0,\n      info: {\n        author: data.info?.Author ?? '',\n        subject: data.info?.Subject ?? '',\n        keywords: data.info?.Keywords ?? '',\n        creator: data.info?.Creator ?? '',\n      },\n      sizeBytes,\n    };\n  } catch (e: any) {\n    console.error(`[pdf-extractor] Failed to parse PDF ${url}: ${e.message}`);\n    return null;\n  }\n}\n\nfunction extractTitleFromUrl(url: string): string {\n  try {\n    const { pathname } = new URL(url);\n    const filename = pathname.split('/').pop() ?? 'document';\n    return filename.replace(/\\.pdf$/i, '').replace(/[-_]/g, ' ').trim();\n  } catch {\n    return 'PDF Document';\n  }\n}\n\nfunction pdfTextToMarkdown(text: string, title: string): string {\n  if (!text.trim()) return `# ${title}\\n\\n*No extractable text content.*`;\n\n  // Clean up common PDF extraction artifacts\n  const cleaned = text\n    .replace(/\\f/g, '\\n\\n---\\n\\n') // form feed = page break\n    .replace(/\\r\\n/g, '\\n')\n    .replace(/\\r/g, '\\n')\n    .replace(/\\n{3,}/g, '\\n\\n')\n    .replace(/[ \\t]{2,}/g, ' ')\n    .trim();\n\n  return `# ${title}\\n\\n${cleaned}`;\n}\n\n/**\n * Filter a list of discovered URLs to find PDF links.\n * Used by http-engine and browser crawler to decide which links to pass to extractPdfFromUrl.\n */\nexport function filterPdfUrls(urls: string[]): string[] {\n  return urls.filter(u => PDF_EXT.test(u));\n}\n"],
+  "mappings": "AAUA,MAAM,UAAU;AAChB,MAAM,WAAW;AAWV,SAAS,SAAS,KAAa,aAA+B;AACnE,SAAO,QAAQ,KAAK,GAAG,MAAM,aAAa,SAAS,QAAQ,KAAK;AAClE;AAEA,eAAsB,kBAAkB,KAAa,OAIjD,CAAC,GAA8B;AACjC,QAAM,EAAE,YAAY,oBAAoB,YAAY,KAAQ,eAAe,KAAK,OAAO,KAAK,IAAI;AAEhG,MAAI;AACJ,MAAI;AAEJ,MAAI;AACF,UAAM,WAAW,MAAM,MAAM,KAAK;AAAA,MAChC,SAAS,EAAE,cAAc,WAAW,UAAU,sBAAsB;AAAA,MACpE,QAAQ,YAAY,QAAQ,SAAS;AAAA,IACvC,CAAC;AACD,QAAI,CAAC,SAAS,GAAI,QAAO;AAEzB,UAAM,cAAc,SAAS,QAAQ,IAAI,cAAc,KAAK;AAC5D,QAAI,CAAC,YAAY,SAAS,KAAK,KAAK,CAAC,QAAQ,KAAK,GAAG,EAAG,QAAO;AAE/D,UAAM,cAAc,MAAM,SAAS,YAAY;AAC/C,gBAAY,YAAY;AACxB,QAAI,YAAY,cAAc;AAC5B,cAAQ,KAAK,kCAAkC,SAAS,sBAAsB,GAAG,EAAE;AACnF,aAAO;AAAA,IACT;AACA,aAAS,OAAO,KAAK,WAAW;AAAA,EAClC,SAAS,GAAQ;AACf,YAAQ,MAAM,uCAAuC,GAAG,KAAK,EAAE,OAAO,EAAE;AACxE,WAAO;AAAA,EACT;AAEA,MAAI;AAKF,UAAM,WAAW,MAAM,OAAO,4BAA4B,EAAE;AAAA,MAC1D,MAAM,OAAO,WAAkB;AAAA,IACjC;AACA,UAAM,QAAQ,SAAS,WAAW;AAClC,UAAM,OAAO,MAAM,MAAM,QAAQ,EAAE,KAAK,EAAE,CAAC;AAE3C,UAAM,QAAS,KAAK,MAAM,SAAoB,oBAAoB,GAAG;AACrE,UAAM,WAAW,kBAAkB,KAAK,QAAQ,IAAI,KAAK;AAEzD,WAAO;AAAA,MACL;AAAA,MACA;AAAA,MACA;AAAA,MACA,WAAW,KAAK,YAAY;AAAA,MAC5B,MAAM;AAAA,QACJ,QAAQ,KAAK,MAAM,UAAU;AAAA,QAC7B,SAAS,KAAK,MAAM,WAAW;AAAA,QAC/B,UAAU,KAAK,MAAM,YAAY;AAAA,QACjC,SAAS,KAAK,MAAM,WAAW;AAAA,MACjC;AAAA,MACA;AAAA,IACF;AAAA,EACF,SAAS,GAAQ;AACf,YAAQ,MAAM,uCAAuC,GAAG,KAAK,EAAE,OAAO,EAAE;AACxE,WAAO;AAAA,EACT;AACF;AAEA,SAAS,oBAAoB,KAAqB;AAChD,MAAI;AACF,UAAM,EAAE,SAAS,IAAI,IAAI,IAAI,GAAG;AAChC,UAAM,WAAW,SAAS,MAAM,GAAG,EAAE,IAAI,KAAK;AAC9C,WAAO,SAAS,QAAQ,WAAW,EAAE,EAAE,QAAQ,SAAS,GAAG,EAAE,KAAK;AAAA,EACpE,QAAQ;AACN,WAAO;AAAA,EACT;AACF;AAEA,SAAS,kBAAkB,MAAc,OAAuB;AAC9D,MAAI,CAAC,KAAK,KAAK,EAAG,QAAO,KAAK,KAAK;AAAA;AAAA;AAGnC,QAAM,UAAU,KACb,QAAQ,OAAO,aAAa,EAC5B,QAAQ,SAAS,IAAI,EACrB,QAAQ,OAAO,IAAI,EACnB,QAAQ,WAAW,MAAM,EACzB,QAAQ,cAAc,GAAG,EACzB,KAAK;AAER,SAAO,KAAK,KAAK;AAAA;AAAA,EAAO,OAAO;AACjC;AAMO,SAAS,cAAc,MAA0B;AACtD,SAAO,KAAK,OAAO,OAAK,QAAQ,KAAK,CAAC,CAAC;AACzC;",
+  "names": []
+}

package/dist/performance-metrics-collector.d.ts

@@ -0,0 +1,50 @@
+/**
+ * Performance metrics collector — captures Core Web Vitals and navigation timing
+ * for pages visited during crawl runs.
+ * Collects LCP, CLS, INP, TTFB, long tasks, and slow resources via PerformanceObserver
+ * and the Navigation/Resource Timing APIs.
+ */
+import type { Page } from 'playwright';
+export type CwvScore = 'good' | 'needs-improvement' | 'poor';
+export interface LcpMetric {
+    value: number;
+    score: CwvScore;
+    elementTag: string | null;
+    url: string | null;
+}
+export interface ClsMetric {
+    value: number;
+    score: CwvScore;
+    shiftCount: number;
+}
+export interface InpMetric {
+    p75: number;
+    score: CwvScore;
+    sampleCount: number;
+}
+export interface NavTimingMetric {
+    ttfb: number;
+    ttfbScore: CwvScore;
+    domContentLoaded: number;
+    domComplete: number;
+    loadComplete: number;
+    transferSizeBytes: number;
+    protocol: string;
+}
+export interface SlowResource {
+    path: string;
+    durationMs: number;
+    transferSizeBytes: number;
+    initiatorType: string;
+}
+export interface PerformanceMetricsResult {
+    lcp: LcpMetric | null;
+    cls: ClsMetric | null;
+    inp: InpMetric | null;
+    navTiming: NavTimingMetric | null;
+    slowResources: SlowResource[];
+    longTaskCount: number;
+    captureSuccess: boolean;
+}
+export declare function injectPerformanceObservers(page: Page): Promise<void>;
+export declare function collectPagePerformance(page: Page): Promise<PerformanceMetricsResult>;

package/dist/performance-metrics-collector.js

@@ -0,0 +1,155 @@
+async function injectPerformanceObservers(page) {
+  await page.addInitScript(() => {
+    window.__zetaPerf = { lcp: null, cls: 0, clsCount: 0, inp: [], longTaskCount: 0 };
+    try {
+      new PerformanceObserver((list) => {
+        const entries = list.getEntries();
+        const last = entries[entries.length - 1];
+        if (last) {
+          window.__zetaPerf.lcp = {
+            value: last.startTime,
+            elementTag: last.element?.tagName || null,
+            url: last.url || null
+          };
+        }
+      }).observe({ entryTypes: ["largest-contentful-paint"] });
+    } catch {
+    }
+    try {
+      new PerformanceObserver((list) => {
+        for (const entry of list.getEntries()) {
+          if (!entry.hadRecentInput) {
+            window.__zetaPerf.cls += entry.value;
+            window.__zetaPerf.clsCount++;
+          }
+        }
+      }).observe({ entryTypes: ["layout-shift"] });
+    } catch {
+    }
+    try {
+      new PerformanceObserver((list) => {
+        for (const entry of list.getEntries()) {
+          window.__zetaPerf.inp.push(entry.duration);
+        }
+      }).observe({ entryTypes: ["event"], durationThreshold: 40 });
+    } catch {
+    }
+    try {
+      new PerformanceObserver((list) => {
+        window.__zetaPerf.longTaskCount += list.getEntries().length;
+      }).observe({ entryTypes: ["longtask"] });
+    } catch {
+    }
+  });
+}
+function scoreLcp(value) {
+  if (value <= 2500) return "good";
+  if (value <= 4e3) return "needs-improvement";
+  return "poor";
+}
+function scoreCls(value) {
+  if (value <= 0.1) return "good";
+  if (value <= 0.25) return "needs-improvement";
+  return "poor";
+}
+function scoreInp(p75) {
+  if (p75 <= 200) return "good";
+  if (p75 <= 500) return "needs-improvement";
+  return "poor";
+}
+function scoreTtfb(value) {
+  if (value <= 800) return "good";
+  if (value <= 1800) return "needs-improvement";
+  return "poor";
+}
+async function collectMetrics(page) {
+  const rawMetrics = await page.evaluate(() => {
+    const p = window.__zetaPerf || {};
+    const nav = performance.getEntriesByType("navigation")[0];
+    const resources = performance.getEntriesByType("resource").filter((r) => r.duration > 500).sort((a, b) => b.duration - a.duration).slice(0, 10).map((r) => ({
+      path: (() => {
+        try {
+          return new URL(r.name).pathname;
+        } catch {
+          return r.name.slice(0, 80);
+        }
+      })(),
+      durationMs: Math.round(r.duration),
+      transferSizeBytes: r.transferSize || 0,
+      initiatorType: r.initiatorType || ""
+    }));
+    const inpValues = p.inp || [];
+    const p75 = inpValues.length > 0 ? inpValues.sort((a, b) => a - b)[Math.floor(inpValues.length * 0.75)] : 0;
+    return {
+      lcp: p.lcp || null,
+      cls: p.cls ?? null,
+      clsCount: p.clsCount ?? 0,
+      inpP75: p75,
+      inpCount: inpValues.length,
+      longTaskCount: p.longTaskCount || 0,
+      nav: nav ? {
+        ttfb: Math.round(nav.responseStart - nav.requestStart),
+        domContentLoaded: Math.round(nav.domContentLoadedEventEnd - nav.startTime),
+        domComplete: Math.round(nav.domComplete - nav.startTime),
+        loadComplete: Math.round(nav.loadEventEnd - nav.startTime),
+        transferSizeBytes: nav.transferSize || 0,
+        protocol: nav.nextHopProtocol || ""
+      } : null,
+      resources
+    };
+  });
+  const lcp = rawMetrics.lcp ? {
+    value: rawMetrics.lcp.value,
+    score: scoreLcp(rawMetrics.lcp.value),
+    elementTag: rawMetrics.lcp.elementTag,
+    url: rawMetrics.lcp.url
+  } : null;
+  const cls = rawMetrics.cls !== null ? {
+    value: rawMetrics.cls,
+    score: scoreCls(rawMetrics.cls),
+    shiftCount: rawMetrics.clsCount
+  } : null;
+  const inp = rawMetrics.inpCount > 0 ? {
+    p75: rawMetrics.inpP75,
+    score: scoreInp(rawMetrics.inpP75),
+    sampleCount: rawMetrics.inpCount
+  } : null;
+  const navTiming = rawMetrics.nav ? {
+    ttfb: rawMetrics.nav.ttfb,
+    ttfbScore: scoreTtfb(rawMetrics.nav.ttfb),
+    domContentLoaded: rawMetrics.nav.domContentLoaded,
+    domComplete: rawMetrics.nav.domComplete,
+    loadComplete: rawMetrics.nav.loadComplete,
+    transferSizeBytes: rawMetrics.nav.transferSizeBytes,
+    protocol: rawMetrics.nav.protocol
+  } : null;
+  return {
+    lcp,
+    cls,
+    inp,
+    navTiming,
+    slowResources: rawMetrics.resources,
+    longTaskCount: rawMetrics.longTaskCount,
+    captureSuccess: true
+  };
+}
+const emptyResult = {
+  lcp: null,
+  cls: null,
+  inp: null,
+  navTiming: null,
+  slowResources: [],
+  longTaskCount: 0
+};
+async function collectPagePerformance(page) {
+  try {
+    return await collectMetrics(page);
+  } catch {
+    return { ...emptyResult, captureSuccess: false };
+  }
+}
+export {
+  collectPagePerformance,
+  injectPerformanceObservers
+};
+//# sourceMappingURL=performance-metrics-collector.js.map

package/dist/performance-metrics-collector.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/performance-metrics-collector.ts"],
+  "sourcesContent": ["/**\n * Performance metrics collector \u2014 captures Core Web Vitals and navigation timing\n * for pages visited during crawl runs.\n * Collects LCP, CLS, INP, TTFB, long tasks, and slow resources via PerformanceObserver\n * and the Navigation/Resource Timing APIs.\n */\n\nimport type { Page } from 'playwright';\n\nexport type CwvScore = 'good' | 'needs-improvement' | 'poor';\n\nexport interface LcpMetric {\n  value: number;\n  score: CwvScore;\n  elementTag: string | null;\n  url: string | null;\n}\n\nexport interface ClsMetric {\n  value: number;\n  score: CwvScore;\n  shiftCount: number;\n}\n\nexport interface InpMetric {\n  p75: number;\n  score: CwvScore;\n  sampleCount: number;\n}\n\nexport interface NavTimingMetric {\n  ttfb: number;\n  ttfbScore: CwvScore;\n  domContentLoaded: number;\n  domComplete: number;\n  loadComplete: number;\n  transferSizeBytes: number;\n  protocol: string;\n}\n\nexport interface SlowResource {\n  path: string;\n  durationMs: number;\n  transferSizeBytes: number;\n  initiatorType: string;\n}\n\nexport interface PerformanceMetricsResult {\n  lcp: LcpMetric | null;\n  cls: ClsMetric | null;\n  inp: InpMetric | null;\n  navTiming: NavTimingMetric | null;\n  slowResources: SlowResource[];\n  longTaskCount: number;\n  captureSuccess: boolean;\n}\n\n// ---------------------------------------------------------------------------\n// Observer injection \u2014 call before page navigation via addInitScript\n// ---------------------------------------------------------------------------\n\nexport async function injectPerformanceObservers(page: Page): Promise<void> {\n  await page.addInitScript(() => {\n    (window as any).__zetaPerf = { lcp: null, cls: 0, clsCount: 0, inp: [], longTaskCount: 0 };\n\n    try {\n      new PerformanceObserver(list => {\n        const entries = list.getEntries();\n        const last = entries[entries.length - 1] as any;\n        if (last) {\n          (window as any).__zetaPerf.lcp = {\n            value: last.startTime,\n            elementTag: last.element?.tagName || null,\n            url: last.url || null,\n          };\n        }\n      }).observe({ entryTypes: ['largest-contentful-paint'] });\n    } catch {}\n\n    try {\n      new PerformanceObserver(list => {\n        for (const entry of list.getEntries() as any[]) {\n          if (!entry.hadRecentInput) {\n            (window as any).__zetaPerf.cls += entry.value;\n            (window as any).__zetaPerf.clsCount++;\n          }\n        }\n      }).observe({ entryTypes: ['layout-shift'] });\n    } catch {}\n\n    try {\n      new PerformanceObserver(list => {\n        for (const entry of list.getEntries() as any[]) {\n          (window as any).__zetaPerf.inp.push(entry.duration);\n        }\n      }).observe({ entryTypes: ['event'], durationThreshold: 40 } as any);\n    } catch {}\n\n    try {\n      new PerformanceObserver(list => {\n        (window as any).__zetaPerf.longTaskCount += list.getEntries().length;\n      }).observe({ entryTypes: ['longtask'] });\n    } catch {}\n  });\n}\n\n// ---------------------------------------------------------------------------\n// Scoring helpers\n// ---------------------------------------------------------------------------\n\nfunction scoreLcp(value: number): CwvScore {\n  if (value <= 2500) return 'good';\n  if (value <= 4000) return 'needs-improvement';\n  return 'poor';\n}\n\nfunction scoreCls(value: number): CwvScore {\n  if (value <= 0.1) return 'good';\n  if (value <= 0.25) return 'needs-improvement';\n  return 'poor';\n}\n\nfunction scoreInp(p75: number): CwvScore {\n  if (p75 <= 200) return 'good';\n  if (p75 <= 500) return 'needs-improvement';\n  return 'poor';\n}\n\nfunction scoreTtfb(value: number): CwvScore {\n  if (value <= 800) return 'good';\n  if (value <= 1800) return 'needs-improvement';\n  return 'poor';\n}\n\n// ---------------------------------------------------------------------------\n// Internal collector\n// ---------------------------------------------------------------------------\n\nasync function collectMetrics(page: Page): Promise<PerformanceMetricsResult> {\n  const rawMetrics = await page.evaluate(() => {\n    const p = (window as any).__zetaPerf || {};\n    const nav = performance.getEntriesByType('navigation')[0] as PerformanceNavigationTiming | undefined;\n    const resources = performance.getEntriesByType('resource')\n      .filter(r => r.duration > 500)\n      .sort((a, b) => b.duration - a.duration)\n      .slice(0, 10)\n      .map(r => ({\n        path: (() => { try { return new URL(r.name).pathname; } catch { return r.name.slice(0, 80); } })(),\n        durationMs: Math.round(r.duration),\n        transferSizeBytes: (r as any).transferSize || 0,\n        initiatorType: (r as any).initiatorType || '',\n      }));\n    const inpValues = p.inp || [];\n    const p75 = inpValues.length > 0\n      ? inpValues.sort((a: number, b: number) => a - b)[Math.floor(inpValues.length * 0.75)]\n      : 0;\n    return {\n      lcp: p.lcp || null,\n      cls: p.cls ?? null,\n      clsCount: p.clsCount ?? 0,\n      inpP75: p75,\n      inpCount: inpValues.length,\n      longTaskCount: p.longTaskCount || 0,\n      nav: nav ? {\n        ttfb: Math.round(nav.responseStart - nav.requestStart),\n        domContentLoaded: Math.round(nav.domContentLoadedEventEnd - nav.startTime),\n        domComplete: Math.round(nav.domComplete - nav.startTime),\n        loadComplete: Math.round(nav.loadEventEnd - nav.startTime),\n        transferSizeBytes: (nav as any).transferSize || 0,\n        protocol: (nav as any).nextHopProtocol || '',\n      } : null,\n      resources,\n    };\n  });\n\n  const lcp: LcpMetric | null = rawMetrics.lcp\n    ? {\n        value: rawMetrics.lcp.value,\n        score: scoreLcp(rawMetrics.lcp.value),\n        elementTag: rawMetrics.lcp.elementTag,\n        url: rawMetrics.lcp.url,\n      }\n    : null;\n\n  const cls: ClsMetric | null = rawMetrics.cls !== null\n    ? {\n        value: rawMetrics.cls,\n        score: scoreCls(rawMetrics.cls),\n        shiftCount: rawMetrics.clsCount,\n      }\n    : null;\n\n  const inp: InpMetric | null = rawMetrics.inpCount > 0\n    ? {\n        p75: rawMetrics.inpP75,\n        score: scoreInp(rawMetrics.inpP75),\n        sampleCount: rawMetrics.inpCount,\n      }\n    : null;\n\n  const navTiming: NavTimingMetric | null = rawMetrics.nav\n    ? {\n        ttfb: rawMetrics.nav.ttfb,\n        ttfbScore: scoreTtfb(rawMetrics.nav.ttfb),\n        domContentLoaded: rawMetrics.nav.domContentLoaded,\n        domComplete: rawMetrics.nav.domComplete,\n        loadComplete: rawMetrics.nav.loadComplete,\n        transferSizeBytes: rawMetrics.nav.transferSizeBytes,\n        protocol: rawMetrics.nav.protocol,\n      }\n    : null;\n\n  return {\n    lcp,\n    cls,\n    inp,\n    navTiming,\n    slowResources: rawMetrics.resources,\n    longTaskCount: rawMetrics.longTaskCount,\n    captureSuccess: true,\n  };\n}\n\n// ---------------------------------------------------------------------------\n// Public API\n// ---------------------------------------------------------------------------\n\nconst emptyResult: Omit<PerformanceMetricsResult, 'captureSuccess'> = {\n  lcp: null,\n  cls: null,\n  inp: null,\n  navTiming: null,\n  slowResources: [],\n  longTaskCount: 0,\n};\n\nexport async function collectPagePerformance(page: Page): Promise<PerformanceMetricsResult> {\n  try {\n    return await collectMetrics(page);\n  } catch {\n    return { ...emptyResult, captureSuccess: false };\n  }\n}\n"],
+  "mappings": "AA6DA,eAAsB,2BAA2B,MAA2B;AAC1E,QAAM,KAAK,cAAc,MAAM;AAC7B,IAAC,OAAe,aAAa,EAAE,KAAK,MAAM,KAAK,GAAG,UAAU,GAAG,KAAK,CAAC,GAAG,eAAe,EAAE;AAEzF,QAAI;AACF,UAAI,oBAAoB,UAAQ;AAC9B,cAAM,UAAU,KAAK,WAAW;AAChC,cAAM,OAAO,QAAQ,QAAQ,SAAS,CAAC;AACvC,YAAI,MAAM;AACR,UAAC,OAAe,WAAW,MAAM;AAAA,YAC/B,OAAO,KAAK;AAAA,YACZ,YAAY,KAAK,SAAS,WAAW;AAAA,YACrC,KAAK,KAAK,OAAO;AAAA,UACnB;AAAA,QACF;AAAA,MACF,CAAC,EAAE,QAAQ,EAAE,YAAY,CAAC,0BAA0B,EAAE,CAAC;AAAA,IACzD,QAAQ;AAAA,IAAC;AAET,QAAI;AACF,UAAI,oBAAoB,UAAQ;AAC9B,mBAAW,SAAS,KAAK,WAAW,GAAY;AAC9C,cAAI,CAAC,MAAM,gBAAgB;AACzB,YAAC,OAAe,WAAW,OAAO,MAAM;AACxC,YAAC,OAAe,WAAW;AAAA,UAC7B;AAAA,QACF;AAAA,MACF,CAAC,EAAE,QAAQ,EAAE,YAAY,CAAC,cAAc,EAAE,CAAC;AAAA,IAC7C,QAAQ;AAAA,IAAC;AAET,QAAI;AACF,UAAI,oBAAoB,UAAQ;AAC9B,mBAAW,SAAS,KAAK,WAAW,GAAY;AAC9C,UAAC,OAAe,WAAW,IAAI,KAAK,MAAM,QAAQ;AAAA,QACpD;AAAA,MACF,CAAC,EAAE,QAAQ,EAAE,YAAY,CAAC,OAAO,GAAG,mBAAmB,GAAG,CAAQ;AAAA,IACpE,QAAQ;AAAA,IAAC;AAET,QAAI;AACF,UAAI,oBAAoB,UAAQ;AAC9B,QAAC,OAAe,WAAW,iBAAiB,KAAK,WAAW,EAAE;AAAA,MAChE,CAAC,EAAE,QAAQ,EAAE,YAAY,CAAC,UAAU,EAAE,CAAC;AAAA,IACzC,QAAQ;AAAA,IAAC;AAAA,EACX,CAAC;AACH;AAMA,SAAS,SAAS,OAAyB;AACzC,MAAI,SAAS,KAAM,QAAO;AAC1B,MAAI,SAAS,IAAM,QAAO;AAC1B,SAAO;AACT;AAEA,SAAS,SAAS,OAAyB;AACzC,MAAI,SAAS,IAAK,QAAO;AACzB,MAAI,SAAS,KAAM,QAAO;AAC1B,SAAO;AACT;AAEA,SAAS,SAAS,KAAuB;AACvC,MAAI,OAAO,IAAK,QAAO;AACvB,MAAI,OAAO,IAAK,QAAO;AACvB,SAAO;AACT;AAEA,SAAS,UAAU,OAAyB;AAC1C,MAAI,SAAS,IAAK,QAAO;AACzB,MAAI,SAAS,KAAM,QAAO;AAC1B,SAAO;AACT;AAMA,eAAe,eAAe,MAA+C;AAC3E,QAAM,aAAa,MAAM,KAAK,SAAS,MAAM;AAC3C,UAAM,IAAK,OAAe,cAAc,CAAC;AACzC,UAAM,MAAM,YAAY,iBAAiB,YAAY,EAAE,CAAC;AACxD,UAAM,YAAY,YAAY,iBAAiB,UAAU,EACtD,OAAO,OAAK,EAAE,WAAW,GAAG,EAC5B,KAAK,CAAC,GAAG,MAAM,EAAE,WAAW,EAAE,QAAQ,EACtC,MAAM,GAAG,EAAE,EACX,IAAI,QAAM;AAAA,MACT,OAAO,MAAM;AAAE,YAAI;AAAE,iBAAO,IAAI,IAAI,EAAE,IAAI,EAAE;AAAA,QAAU,QAAQ;AAAE,iBAAO,EAAE,KAAK,MAAM,GAAG,EAAE;AAAA,QAAG;AAAA,MAAE,GAAG;AAAA,MACjG,YAAY,KAAK,MAAM,EAAE,QAAQ;AAAA,MACjC,mBAAoB,EAAU,gBAAgB;AAAA,MAC9C,eAAgB,EAAU,iBAAiB;AAAA,IAC7C,EAAE;AACJ,UAAM,YAAY,EAAE,OAAO,CAAC;AAC5B,UAAM,MAAM,UAAU,SAAS,IAC3B,UAAU,KAAK,CAAC,GAAW,MAAc,IAAI,CAAC,EAAE,KAAK,MAAM,UAAU,SAAS,IAAI,CAAC,IACnF;AACJ,WAAO;AAAA,MACL,KAAK,EAAE,OAAO;AAAA,MACd,KAAK,EAAE,OAAO;AAAA,MACd,UAAU,EAAE,YAAY;AAAA,MACxB,QAAQ;AAAA,MACR,UAAU,UAAU;AAAA,MACpB,eAAe,EAAE,iBAAiB;AAAA,MAClC,KAAK,MAAM;AAAA,QACT,MAAM,KAAK,MAAM,IAAI,gBAAgB,IAAI,YAAY;AAAA,QACrD,kBAAkB,KAAK,MAAM,IAAI,2BAA2B,IAAI,SAAS;AAAA,QACzE,aAAa,KAAK,MAAM,IAAI,cAAc,IAAI,SAAS;AAAA,QACvD,cAAc,KAAK,MAAM,IAAI,eAAe,IAAI,SAAS;AAAA,QACzD,mBAAoB,IAAY,gBAAgB;AAAA,QAChD,UAAW,IAAY,mBAAmB;AAAA,MAC5C,IAAI;AAAA,MACJ;AAAA,IACF;AAAA,EACF,CAAC;AAED,QAAM,MAAwB,WAAW,MACrC;AAAA,IACE,OAAO,WAAW,IAAI;AAAA,IACtB,OAAO,SAAS,WAAW,IAAI,KAAK;AAAA,IACpC,YAAY,WAAW,IAAI;AAAA,IAC3B,KAAK,WAAW,IAAI;AAAA,EACtB,IACA;AAEJ,QAAM,MAAwB,WAAW,QAAQ,OAC7C;AAAA,IACE,OAAO,WAAW;AAAA,IAClB,OAAO,SAAS,WAAW,GAAG;AAAA,IAC9B,YAAY,WAAW;AAAA,EACzB,IACA;AAEJ,QAAM,MAAwB,WAAW,WAAW,IAChD;AAAA,IACE,KAAK,WAAW;AAAA,IAChB,OAAO,SAAS,WAAW,MAAM;AAAA,IACjC,aAAa,WAAW;AAAA,EAC1B,IACA;AAEJ,QAAM,YAAoC,WAAW,MACjD;AAAA,IACE,MAAM,WAAW,IAAI;AAAA,IACrB,WAAW,UAAU,WAAW,IAAI,IAAI;AAAA,IACxC,kBAAkB,WAAW,IAAI;AAAA,IACjC,aAAa,WAAW,IAAI;AAAA,IAC5B,cAAc,WAAW,IAAI;AAAA,IAC7B,mBAAmB,WAAW,IAAI;AAAA,IAClC,UAAU,WAAW,IAAI;AAAA,EAC3B,IACA;AAEJ,SAAO;AAAA,IACL;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,IACA,eAAe,WAAW;AAAA,IAC1B,eAAe,WAAW;AAAA,IAC1B,gBAAgB;AAAA,EAClB;AACF;AAMA,MAAM,cAAgE;AAAA,EACpE,KAAK;AAAA,EACL,KAAK;AAAA,EACL,KAAK;AAAA,EACL,WAAW;AAAA,EACX,eAAe,CAAC;AAAA,EAChB,eAAe;AACjB;AAEA,eAAsB,uBAAuB,MAA+C;AAC1F,MAAI;AACF,WAAO,MAAM,eAAe,IAAI;AAAA,EAClC,QAAQ;AACN,WAAO,EAAE,GAAG,aAAa,gBAAgB,MAAM;AAAA,EACjD;AACF;",
+  "names": []
+}

package/dist/performance-metrics.d.ts

@@ -0,0 +1,52 @@
+/**
+ * Performance metrics collection via CDP + Playwright page.evaluate().
+ *
+ * Collects Core Web Vitals and resource timing after page navigation.
+ * Stored alongside screen data for test generation context and regression detection.
+ *
+ * Metrics:
+ * - LCP (Largest Contentful Paint) — via PerformanceObserver in page
+ * - CLS (Cumulative Layout Shift) — via PerformanceObserver in page
+ * - TTFB (Time to First Byte) — from navigation timing
+ * - DOMContentLoaded — from navigation timing
+ * - Resource count and size by type
+ */
+export interface CoreWebVitals {
+    lcp?: number;
+    cls?: number;
+    ttfb?: number;
+    fcp?: number;
+    domContentLoaded?: number;
+    loadEventEnd?: number;
+}
+export interface ResourceSummary {
+    scripts: {
+        count: number;
+        totalBytes: number;
+    };
+    styles: {
+        count: number;
+        totalBytes: number;
+    };
+    images: {
+        count: number;
+        totalBytes: number;
+    };
+    fonts: {
+        count: number;
+        totalBytes: number;
+    };
+    other: {
+        count: number;
+        totalBytes: number;
+    };
+}
+export interface PagePerfMetrics {
+    url: string;
+    vitals: CoreWebVitals;
+    resources: ResourceSummary;
+    totalTransferBytes: number;
+    requestCount: number;
+    measuredAt: string;
+}
+export declare function collectPerfMetrics(page: any, url: string): Promise<PagePerfMetrics>;

package/dist/performance-metrics.js

@@ -0,0 +1,83 @@
+async function collectPerfMetrics(page, url) {
+  const measuredAt = (/* @__PURE__ */ new Date()).toISOString();
+  let vitals = {};
+  try {
+    vitals = await page.evaluate(() => {
+      return new Promise((resolve) => {
+        const result = {};
+        const nav = performance.getEntriesByType("navigation")[0];
+        if (nav) {
+          result.ttfb = Math.round(nav.responseStart - nav.requestStart);
+          result.domContentLoaded = Math.round(nav.domContentLoadedEventEnd);
+          result.loadEventEnd = Math.round(nav.loadEventEnd);
+        }
+        let lcpDone = false;
+        let clsDone = false;
+        const tryResolve = () => {
+          if (lcpDone && clsDone) resolve(result);
+        };
+        try {
+          new PerformanceObserver((list) => {
+            const entries = list.getEntries();
+            if (entries.length > 0) result.lcp = Math.round(entries[entries.length - 1].startTime);
+            lcpDone = true;
+            tryResolve();
+          }).observe({ type: "largest-contentful-paint", buffered: true });
+        } catch {
+          lcpDone = true;
+        }
+        try {
+          let cls = 0;
+          new PerformanceObserver((list) => {
+            for (const e of list.getEntries()) {
+              if (!e.hadRecentInput) cls += e.value;
+            }
+            result.cls = Math.round(cls * 1e3) / 1e3;
+            clsDone = true;
+            tryResolve();
+          }).observe({ type: "layout-shift", buffered: true });
+        } catch {
+          clsDone = true;
+        }
+        try {
+          const fcp = performance.getEntriesByName("first-contentful-paint")[0];
+          if (fcp) result.fcp = Math.round(fcp.startTime);
+        } catch {
+        }
+        setTimeout(() => resolve(result), 3e3);
+      });
+    }).catch(() => ({}));
+  } catch {
+  }
+  let resources = {
+    scripts: { count: 0, totalBytes: 0 },
+    styles: { count: 0, totalBytes: 0 },
+    images: { count: 0, totalBytes: 0 },
+    fonts: { count: 0, totalBytes: 0 },
+    other: { count: 0, totalBytes: 0 }
+  };
+  let totalTransferBytes = 0;
+  let requestCount = 0;
+  try {
+    const entries = await page.evaluate(
+      () => performance.getEntriesByType("resource").map((e) => ({
+        name: e.name,
+        initiatorType: e.initiatorType,
+        transferSize: e.transferSize ?? 0
+      }))
+    ).catch(() => []);
+    for (const e of entries) {
+      requestCount++;
+      totalTransferBytes += e.transferSize;
+      const cat = e.initiatorType === "script" ? "scripts" : e.initiatorType === "css" || e.name?.endsWith(".css") ? "styles" : e.initiatorType === "img" ? "images" : e.name?.match(/\.(woff|woff2|ttf|otf|eot)/) ? "fonts" : "other";
+      resources[cat].count++;
+      resources[cat].totalBytes += e.transferSize;
+    }
+  } catch {
+  }
+  return { url, vitals, resources, totalTransferBytes, requestCount, measuredAt };
+}
+export {
+  collectPerfMetrics
+};
+//# sourceMappingURL=performance-metrics.js.map

package/dist/performance-metrics.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/performance-metrics.ts"],
+  "sourcesContent": ["/**\n * Performance metrics collection via CDP + Playwright page.evaluate().\n *\n * Collects Core Web Vitals and resource timing after page navigation.\n * Stored alongside screen data for test generation context and regression detection.\n *\n * Metrics:\n * - LCP (Largest Contentful Paint) \u2014 via PerformanceObserver in page\n * - CLS (Cumulative Layout Shift) \u2014 via PerformanceObserver in page\n * - TTFB (Time to First Byte) \u2014 from navigation timing\n * - DOMContentLoaded \u2014 from navigation timing\n * - Resource count and size by type\n */\n\nimport type { Page } from 'playwright';\n\nexport interface CoreWebVitals {\n  lcp?: number;   // ms\n  cls?: number;   // unitless score\n  ttfb?: number;  // ms\n  fcp?: number;   // First Contentful Paint ms\n  domContentLoaded?: number; // ms\n  loadEventEnd?: number;     // ms\n}\n\nexport interface ResourceSummary {\n  scripts: { count: number; totalBytes: number };\n  styles: { count: number; totalBytes: number };\n  images: { count: number; totalBytes: number };\n  fonts: { count: number; totalBytes: number };\n  other: { count: number; totalBytes: number };\n}\n\nexport interface PagePerfMetrics {\n  url: string;\n  vitals: CoreWebVitals;\n  resources: ResourceSummary;\n  totalTransferBytes: number;\n  requestCount: number;\n  measuredAt: string;\n}\n\nexport async function collectPerfMetrics(page: any, url: string): Promise<PagePerfMetrics> {\n  const measuredAt = new Date().toISOString();\n\n  // Inject PerformanceObserver for LCP + CLS, wait up to 3s\n  let vitals: CoreWebVitals = {};\n  try {\n    vitals = await page.evaluate(() => {\n      return new Promise<any>((resolve) => {\n        const result: any = {};\n\n        // Navigation timing\n        const nav = performance.getEntriesByType('navigation')[0] as any;\n        if (nav) {\n          result.ttfb = Math.round(nav.responseStart - nav.requestStart);\n          result.domContentLoaded = Math.round(nav.domContentLoadedEventEnd);\n          result.loadEventEnd = Math.round(nav.loadEventEnd);\n        }\n\n        let lcpDone = false;\n        let clsDone = false;\n        const tryResolve = () => { if (lcpDone && clsDone) resolve(result); };\n\n        // LCP\n        try {\n          new PerformanceObserver((list) => {\n            const entries = list.getEntries();\n            if (entries.length > 0) result.lcp = Math.round((entries[entries.length - 1] as any).startTime);\n            lcpDone = true;\n            tryResolve();\n          }).observe({ type: 'largest-contentful-paint', buffered: true });\n        } catch { lcpDone = true; }\n\n        // CLS\n        try {\n          let cls = 0;\n          new PerformanceObserver((list) => {\n            for (const e of list.getEntries() as any[]) {\n              if (!e.hadRecentInput) cls += e.value;\n            }\n            result.cls = Math.round(cls * 1000) / 1000;\n            clsDone = true;\n            tryResolve();\n          }).observe({ type: 'layout-shift', buffered: true });\n        } catch { clsDone = true; }\n\n        // FCP\n        try {\n          const fcp = performance.getEntriesByName('first-contentful-paint')[0] as any;\n          if (fcp) result.fcp = Math.round(fcp.startTime);\n        } catch {}\n\n        // Fallback resolve after 3s\n        setTimeout(() => resolve(result), 3_000);\n      });\n    }).catch(() => ({}));\n  } catch { /* page closed or navigate already */ }\n\n  // Resource timing\n  let resources: ResourceSummary = {\n    scripts: { count: 0, totalBytes: 0 },\n    styles: { count: 0, totalBytes: 0 },\n    images: { count: 0, totalBytes: 0 },\n    fonts: { count: 0, totalBytes: 0 },\n    other: { count: 0, totalBytes: 0 },\n  };\n  let totalTransferBytes = 0;\n  let requestCount = 0;\n\n  try {\n    const entries = await page.evaluate(() =>\n      performance.getEntriesByType('resource').map((e: any) => ({\n        name: e.name,\n        initiatorType: e.initiatorType,\n        transferSize: e.transferSize ?? 0,\n      }))\n    ).catch(() => [] as any[]);\n\n    for (const e of entries) {\n      requestCount++;\n      totalTransferBytes += e.transferSize;\n      const cat = e.initiatorType === 'script' ? 'scripts'\n        : e.initiatorType === 'css' || e.name?.endsWith('.css') ? 'styles'\n        : e.initiatorType === 'img' ? 'images'\n        : e.name?.match(/\\.(woff|woff2|ttf|otf|eot)/) ? 'fonts'\n        : 'other';\n      resources[cat as keyof ResourceSummary].count++;\n      resources[cat as keyof ResourceSummary].totalBytes += e.transferSize;\n    }\n  } catch {}\n\n  return { url, vitals, resources, totalTransferBytes, requestCount, measuredAt };\n}\n"],
+  "mappings": "AA0CA,eAAsB,mBAAmB,MAAW,KAAuC;AACzF,QAAM,cAAa,oBAAI,KAAK,GAAE,YAAY;AAG1C,MAAI,SAAwB,CAAC;AAC7B,MAAI;AACF,aAAS,MAAM,KAAK,SAAS,MAAM;AACjC,aAAO,IAAI,QAAa,CAAC,YAAY;AACnC,cAAM,SAAc,CAAC;AAGrB,cAAM,MAAM,YAAY,iBAAiB,YAAY,EAAE,CAAC;AACxD,YAAI,KAAK;AACP,iBAAO,OAAO,KAAK,MAAM,IAAI,gBAAgB,IAAI,YAAY;AAC7D,iBAAO,mBAAmB,KAAK,MAAM,IAAI,wBAAwB;AACjE,iBAAO,eAAe,KAAK,MAAM,IAAI,YAAY;AAAA,QACnD;AAEA,YAAI,UAAU;AACd,YAAI,UAAU;AACd,cAAM,aAAa,MAAM;AAAE,cAAI,WAAW,QAAS,SAAQ,MAAM;AAAA,QAAG;AAGpE,YAAI;AACF,cAAI,oBAAoB,CAAC,SAAS;AAChC,kBAAM,UAAU,KAAK,WAAW;AAChC,gBAAI,QAAQ,SAAS,EAAG,QAAO,MAAM,KAAK,MAAO,QAAQ,QAAQ,SAAS,CAAC,EAAU,SAAS;AAC9F,sBAAU;AACV,uBAAW;AAAA,UACb,CAAC,EAAE,QAAQ,EAAE,MAAM,4BAA4B,UAAU,KAAK,CAAC;AAAA,QACjE,QAAQ;AAAE,oBAAU;AAAA,QAAM;AAG1B,YAAI;AACF,cAAI,MAAM;AACV,cAAI,oBAAoB,CAAC,SAAS;AAChC,uBAAW,KAAK,KAAK,WAAW,GAAY;AAC1C,kBAAI,CAAC,EAAE,eAAgB,QAAO,EAAE;AAAA,YAClC;AACA,mBAAO,MAAM,KAAK,MAAM,MAAM,GAAI,IAAI;AACtC,sBAAU;AACV,uBAAW;AAAA,UACb,CAAC,EAAE,QAAQ,EAAE,MAAM,gBAAgB,UAAU,KAAK,CAAC;AAAA,QACrD,QAAQ;AAAE,oBAAU;AAAA,QAAM;AAG1B,YAAI;AACF,gBAAM,MAAM,YAAY,iBAAiB,wBAAwB,EAAE,CAAC;AACpE,cAAI,IAAK,QAAO,MAAM,KAAK,MAAM,IAAI,SAAS;AAAA,QAChD,QAAQ;AAAA,QAAC;AAGT,mBAAW,MAAM,QAAQ,MAAM,GAAG,GAAK;AAAA,MACzC,CAAC;AAAA,IACH,CAAC,EAAE,MAAM,OAAO,CAAC,EAAE;AAAA,EACrB,QAAQ;AAAA,EAAwC;AAGhD,MAAI,YAA6B;AAAA,IAC/B,SAAS,EAAE,OAAO,GAAG,YAAY,EAAE;AAAA,IACnC,QAAQ,EAAE,OAAO,GAAG,YAAY,EAAE;AAAA,IAClC,QAAQ,EAAE,OAAO,GAAG,YAAY,EAAE;AAAA,IAClC,OAAO,EAAE,OAAO,GAAG,YAAY,EAAE;AAAA,IACjC,OAAO,EAAE,OAAO,GAAG,YAAY,EAAE;AAAA,EACnC;AACA,MAAI,qBAAqB;AACzB,MAAI,eAAe;AAEnB,MAAI;AACF,UAAM,UAAU,MAAM,KAAK;AAAA,MAAS,MAClC,YAAY,iBAAiB,UAAU,EAAE,IAAI,CAAC,OAAY;AAAA,QACxD,MAAM,EAAE;AAAA,QACR,eAAe,EAAE;AAAA,QACjB,cAAc,EAAE,gBAAgB;AAAA,MAClC,EAAE;AAAA,IACJ,EAAE,MAAM,MAAM,CAAC,CAAU;AAEzB,eAAW,KAAK,SAAS;AACvB;AACA,4BAAsB,EAAE;AACxB,YAAM,MAAM,EAAE,kBAAkB,WAAW,YACvC,EAAE,kBAAkB,SAAS,EAAE,MAAM,SAAS,MAAM,IAAI,WACxD,EAAE,kBAAkB,QAAQ,WAC5B,EAAE,MAAM,MAAM,4BAA4B,IAAI,UAC9C;AACJ,gBAAU,GAA4B,EAAE;AACxC,gBAAU,GAA4B,EAAE,cAAc,EAAE;AAAA,IAC1D;AAAA,EACF,QAAQ;AAAA,EAAC;AAET,SAAO,EAAE,KAAK,QAAQ,WAAW,oBAAoB,cAAc,WAAW;AAChF;",
+  "names": []
+}