@taodigital/crawler 0.0.0-stage → 1.2.0

package/dist/extractors/summary-extractor.js

@@ -0,0 +1,36 @@
+async function extractSummary(markdown, options) {
+  const input = markdown.slice(0, 8e3);
+  try {
+    const { ask } = await import("@taodigital/agents").catch(() => ({ ask: null }));
+    if (!ask) {
+      return makeTruncated(input);
+    }
+    const summary = await ask(
+      "Summarize the following web page content in 2-4 concise sentences. Focus on the main purpose, key features, and primary value proposition. Return only the summary text, no preamble.",
+      input,
+      options?.model ? { provider: "anthropic", model: options.model } : void 0,
+      options?.maxTokens ?? 300
+    );
+    return {
+      summary: summary.trim(),
+      wordCount: markdown.split(/\s+/).length,
+      method: "llm",
+      model: options?.model
+    };
+  } catch (err) {
+    console.warn("[summary-extractor] LLM call failed, using truncated fallback:", err);
+    return makeTruncated(input);
+  }
+}
+function makeTruncated(input) {
+  const excerpt = input.slice(0, 500).trim();
+  return {
+    summary: excerpt + (input.length > 500 ? "..." : ""),
+    wordCount: excerpt.split(/\s+/).length,
+    method: "truncated"
+  };
+}
+export {
+  extractSummary
+};
+//# sourceMappingURL=summary-extractor.js.map

package/dist/extractors/summary-extractor.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../../src/extractors/summary-extractor.ts"],
+  "sourcesContent": ["export interface SummaryResult {\n  summary: string;\n  wordCount: number;\n  method: 'llm' | 'truncated';\n  model?: string;\n}\n\nexport async function extractSummary(\n  markdown: string,\n  options?: { model?: string; maxTokens?: number },\n): Promise<SummaryResult> {\n  // Truncate to 8000 chars to stay within token budget\n  const input = markdown.slice(0, 8000);\n\n  try {\n    // Dynamic import to avoid hard dependency if agents package unavailable\n    const { ask } = await import('@detiq/agents').catch(() => ({ ask: null as any }));\n\n    if (!ask) {\n      return makeTruncated(input);\n    }\n\n    const summary: string = await ask(\n      'Summarize the following web page content in 2-4 concise sentences. Focus on the main purpose, key features, and primary value proposition. Return only the summary text, no preamble.',\n      input,\n      options?.model ? { provider: 'anthropic', model: options.model } : undefined,\n      options?.maxTokens ?? 300,\n    );\n\n    return {\n      summary: summary.trim(),\n      wordCount: markdown.split(/\\s+/).length,\n      method: 'llm',\n      model: options?.model,\n    };\n  } catch (err) {\n    console.warn('[summary-extractor] LLM call failed, using truncated fallback:', err);\n    return makeTruncated(input);\n  }\n}\n\nfunction makeTruncated(input: string): SummaryResult {\n  const excerpt = input.slice(0, 500).trim();\n  return {\n    summary: excerpt + (input.length > 500 ? '...' : ''),\n    wordCount: excerpt.split(/\\s+/).length,\n    method: 'truncated',\n  };\n}\n"],
+  "mappings": "AAOA,eAAsB,eACpB,UACA,SACwB;AAExB,QAAM,QAAQ,SAAS,MAAM,GAAG,GAAI;AAEpC,MAAI;AAEF,UAAM,EAAE,IAAI,IAAI,MAAM,OAAO,eAAe,EAAE,MAAM,OAAO,EAAE,KAAK,KAAY,EAAE;AAEhF,QAAI,CAAC,KAAK;AACR,aAAO,cAAc,KAAK;AAAA,IAC5B;AAEA,UAAM,UAAkB,MAAM;AAAA,MAC5B;AAAA,MACA;AAAA,MACA,SAAS,QAAQ,EAAE,UAAU,aAAa,OAAO,QAAQ,MAAM,IAAI;AAAA,MACnE,SAAS,aAAa;AAAA,IACxB;AAEA,WAAO;AAAA,MACL,SAAS,QAAQ,KAAK;AAAA,MACtB,WAAW,SAAS,MAAM,KAAK,EAAE;AAAA,MACjC,QAAQ;AAAA,MACR,OAAO,SAAS;AAAA,IAClB;AAAA,EACF,SAAS,KAAK;AACZ,YAAQ,KAAK,kEAAkE,GAAG;AAClF,WAAO,cAAc,KAAK;AAAA,EAC5B;AACF;AAEA,SAAS,cAAc,OAA8B;AACnD,QAAM,UAAU,MAAM,MAAM,GAAG,GAAG,EAAE,KAAK;AACzC,SAAO;AAAA,IACL,SAAS,WAAW,MAAM,SAAS,MAAM,QAAQ;AAAA,IACjD,WAAW,QAAQ,MAAM,KAAK,EAAE;AAAA,IAChC,QAAQ;AAAA,EACV;AACF;",
+  "names": []
+}

package/dist/failure-analysis-agent.d.ts

@@ -0,0 +1,16 @@
+export interface FailureContext {
+    tenantId?: string;
+    error: string;
+    stepDescription?: string;
+    consoleLogs?: Array<{
+        type: string;
+        text: string;
+    }>;
+    networkErrors?: Array<{
+        url: string;
+        status: number;
+        method: string;
+    }>;
+    screenshotPath?: string;
+}
+export declare function analyzeFailure(ctx: FailureContext): Promise<string | null>;

package/dist/failure-analysis-agent.js

@@ -0,0 +1,42 @@
+import { readFileSync } from "node:fs";
+import { askTrackedWithImages, askTracked, deterministicSeed } from "@taodigital/agents";
+import { AppBrain } from "@taodigital/app-brain";
+async function analyzeFailure(ctx) {
+  try {
+    let llmCfg;
+    if (ctx.tenantId) {
+      llmCfg = await AppBrain.getTaskLLMConfig(ctx.tenantId, "testRunner").catch(() => void 0) ?? void 0;
+    }
+    if (!llmCfg) return null;
+    const consoleSection = ctx.consoleLogs?.length ? `
+
+Browser console (last ${ctx.consoleLogs.length} entries):
+${ctx.consoleLogs.map((l) => `[${l.type}] ${l.text}`).join("\n")}` : "";
+    const networkSection = ctx.networkErrors?.length ? `
+
+Failed network requests:
+${ctx.networkErrors.map((n) => `${n.method} ${n.url} \u2192 ${n.status}`).join("\n")}` : "";
+    const userText = `A Playwright automated test failed. Identify the root cause in 2-3 sentences. Be specific \u2014 name the failing element, API endpoint, or error type. Do not suggest fixes.
+
+Test step: ${ctx.stepDescription ?? "unknown"}
+Error: ${ctx.error}${consoleSection}${networkSection}`;
+    if (ctx.screenshotPath) {
+      try {
+        const img = readFileSync(ctx.screenshotPath);
+        const imgSeed = deterministicSeed(`failure-img:${ctx.tenantId ?? ""}:${(ctx.stepDescription ?? "").slice(0, 100)}:${ctx.error.slice(0, 200)}`);
+        const result2 = await askTrackedWithImages("", userText, [{ base64: img.toString("base64"), mediaType: "image/png" }], llmCfg, { temperature: 0, seed: imgSeed });
+        return result2.text || null;
+      } catch {
+      }
+    }
+    const failureSeed = deterministicSeed(`${ctx.tenantId ?? ""}:failure:${(ctx.stepDescription ?? "").slice(0, 100)}:${ctx.error.slice(0, 200)}`);
+    const result = await askTracked("", userText, llmCfg, 256, void 0, { temperature: 0, seed: failureSeed });
+    return result.text || null;
+  } catch {
+    return null;
+  }
+}
+export {
+  analyzeFailure
+};
+//# sourceMappingURL=failure-analysis-agent.js.map

package/dist/failure-analysis-agent.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/failure-analysis-agent.ts"],
+  "sourcesContent": ["import { readFileSync } from 'node:fs';\nimport { askTrackedWithImages, askTracked, deterministicSeed } from '@detiq/agents';\nimport type { LLMConfig } from '@detiq/agents';\nimport { AppBrain } from '@detiq/app-brain';\n\nexport interface FailureContext {\n  tenantId?: string;\n  error: string;\n  stepDescription?: string;\n  consoleLogs?: Array<{ type: string; text: string }>;\n  networkErrors?: Array<{ url: string; status: number; method: string }>;\n  screenshotPath?: string;\n}\n\nexport async function analyzeFailure(ctx: FailureContext): Promise<string | null> {\n  try {\n    let llmCfg: LLMConfig | undefined;\n    if (ctx.tenantId) {\n      llmCfg = await AppBrain.getTaskLLMConfig(ctx.tenantId, 'testRunner').catch(() => undefined) as LLMConfig | undefined ?? undefined;\n    }\n    if (!llmCfg) return null;\n\n    const consoleSection = ctx.consoleLogs?.length\n      ? `\\n\\nBrowser console (last ${ctx.consoleLogs.length} entries):\\n${ctx.consoleLogs.map(l => `[${l.type}] ${l.text}`).join('\\n')}`\n      : '';\n\n    const networkSection = ctx.networkErrors?.length\n      ? `\\n\\nFailed network requests:\\n${ctx.networkErrors.map(n => `${n.method} ${n.url} \u2192 ${n.status}`).join('\\n')}`\n      : '';\n\n    const userText = `A Playwright automated test failed. Identify the root cause in 2-3 sentences. Be specific \u2014 name the failing element, API endpoint, or error type. Do not suggest fixes.\n\nTest step: ${ctx.stepDescription ?? 'unknown'}\nError: ${ctx.error}${consoleSection}${networkSection}`;\n\n    if (ctx.screenshotPath) {\n      try {\n        const img = readFileSync(ctx.screenshotPath);\n        const imgSeed = deterministicSeed(`failure-img:${ctx.tenantId ?? ''}:${(ctx.stepDescription ?? '').slice(0, 100)}:${ctx.error.slice(0, 200)}`);\n        const result = await askTrackedWithImages('', userText, [{ base64: img.toString('base64'), mediaType: 'image/png' }], llmCfg, { temperature: 0, seed: imgSeed });\n        return result.text || null;\n      } catch { /* fall through to text-only */ }\n    }\n\n    const failureSeed = deterministicSeed(`${ctx.tenantId ?? ''}:failure:${(ctx.stepDescription ?? '').slice(0, 100)}:${ctx.error.slice(0, 200)}`);\n    const result = await askTracked('', userText, llmCfg, 256, undefined, { temperature: 0, seed: failureSeed });\n    return result.text || null;\n  } catch {\n    return null;\n  }\n}\n"],
+  "mappings": "AAAA,SAAS,oBAAoB;AAC7B,SAAS,sBAAsB,YAAY,yBAAyB;AAEpE,SAAS,gBAAgB;AAWzB,eAAsB,eAAe,KAA6C;AAChF,MAAI;AACF,QAAI;AACJ,QAAI,IAAI,UAAU;AAChB,eAAS,MAAM,SAAS,iBAAiB,IAAI,UAAU,YAAY,EAAE,MAAM,MAAM,MAAS,KAA8B;AAAA,IAC1H;AACA,QAAI,CAAC,OAAQ,QAAO;AAEpB,UAAM,iBAAiB,IAAI,aAAa,SACpC;AAAA;AAAA,wBAA6B,IAAI,YAAY,MAAM;AAAA,EAAe,IAAI,YAAY,IAAI,OAAK,IAAI,EAAE,IAAI,KAAK,EAAE,IAAI,EAAE,EAAE,KAAK,IAAI,CAAC,KAC9H;AAEJ,UAAM,iBAAiB,IAAI,eAAe,SACtC;AAAA;AAAA;AAAA,EAAiC,IAAI,cAAc,IAAI,OAAK,GAAG,EAAE,MAAM,IAAI,EAAE,GAAG,WAAM,EAAE,MAAM,EAAE,EAAE,KAAK,IAAI,CAAC,KAC5G;AAEJ,UAAM,WAAW;AAAA;AAAA,aAER,IAAI,mBAAmB,SAAS;AAAA,SACpC,IAAI,KAAK,GAAG,cAAc,GAAG,cAAc;AAEhD,QAAI,IAAI,gBAAgB;AACtB,UAAI;AACF,cAAM,MAAM,aAAa,IAAI,cAAc;AAC3C,cAAM,UAAU,kBAAkB,eAAe,IAAI,YAAY,EAAE,KAAK,IAAI,mBAAmB,IAAI,MAAM,GAAG,GAAG,CAAC,IAAI,IAAI,MAAM,MAAM,GAAG,GAAG,CAAC,EAAE;AAC7I,cAAMA,UAAS,MAAM,qBAAqB,IAAI,UAAU,CAAC,EAAE,QAAQ,IAAI,SAAS,QAAQ,GAAG,WAAW,YAAY,CAAC,GAAG,QAAQ,EAAE,aAAa,GAAG,MAAM,QAAQ,CAAC;AAC/J,eAAOA,QAAO,QAAQ;AAAA,MACxB,QAAQ;AAAA,MAAkC;AAAA,IAC5C;AAEA,UAAM,cAAc,kBAAkB,GAAG,IAAI,YAAY,EAAE,aAAa,IAAI,mBAAmB,IAAI,MAAM,GAAG,GAAG,CAAC,IAAI,IAAI,MAAM,MAAM,GAAG,GAAG,CAAC,EAAE;AAC7I,UAAM,SAAS,MAAM,WAAW,IAAI,UAAU,QAAQ,KAAK,QAAW,EAAE,aAAa,GAAG,MAAM,YAAY,CAAC;AAC3G,WAAO,OAAO,QAAQ;AAAA,EACxB,QAAQ;AACN,WAAO;AAAA,EACT;AACF;",
+  "names": ["result"]
+}

package/dist/feature-flag-detector.d.ts

@@ -0,0 +1,30 @@
+/**
+ * Feature flag + A/B test detector.
+ *
+ * Detects popular feature flag / experiment SDKs loaded on the page:
+ * LaunchDarkly, Optimizely, Split.io, Unleash, GrowthBook, Statsig, Flagsmith, Amplitude Experiment.
+ *
+ * Also detects common A/B test patterns: data-testid attributes, gtm.js experiment layers,
+ * Optimizely variation IDs, VWO, etc.
+ *
+ * Why: ZeTa needs to know when a page shows different content to different users
+ * so test generation can account for variants and not generate flaky tests.
+ */
+import type { Page } from 'playwright';
+export interface FeatureFlagInfo {
+    sdk: string;
+    confidence: 'high' | 'medium' | 'low';
+    details?: string;
+}
+export interface AbTestInfo {
+    tool: string;
+    experiments?: string[];
+}
+export interface FeatureFlagDetectionResult {
+    url: string;
+    flagSdks: FeatureFlagInfo[];
+    abTools: AbTestInfo[];
+    hasAnyFlags: boolean;
+    detectedAt: string;
+}
+export declare function detectFeatureFlags(page: Page, url: string): Promise<FeatureFlagDetectionResult>;

package/dist/feature-flag-detector.js

@@ -0,0 +1,55 @@
+const FLAG_SDK_CHECKS = [
+  { sdk: "LaunchDarkly", windowKeys: ["LDClient", "LaunchDarkly", "ldClient"], scriptPattern: /launchdarkly/i },
+  { sdk: "Optimizely", windowKeys: ["optimizely", "optimizelySdk"], scriptPattern: /optimizely/i },
+  { sdk: "Split.io", windowKeys: ["splitio", "SplitFactory"], scriptPattern: /split\.io/i },
+  { sdk: "GrowthBook", windowKeys: ["growthbook", "GrowthBook"], scriptPattern: /growthbook/i },
+  { sdk: "Statsig", windowKeys: ["statsig", "StatsigClient"], scriptPattern: /statsig/i },
+  { sdk: "Flagsmith", windowKeys: ["flagsmith"], scriptPattern: /flagsmith/i },
+  { sdk: "Unleash", windowKeys: ["unleash", "UnleashClient"], scriptPattern: /unleash/i },
+  { sdk: "VWO", windowKeys: ["vwo", "_vwo"], scriptPattern: /vwo\.com/i },
+  { sdk: "Google Optimize", windowKeys: ["gaData", "google_optimize"], scriptPattern: /googleoptimize/i },
+  { sdk: "AB Tasty", windowKeys: ["ABTasty"], scriptPattern: /abtasty/i }
+];
+async function detectFeatureFlags(page, url) {
+  const detectedAt = (/* @__PURE__ */ new Date()).toISOString();
+  const result = await page.evaluate((checks) => {
+    const flagSdks = [];
+    const scripts = Array.from(document.querySelectorAll("script[src]")).map((s) => s.src ?? "");
+    for (const check of checks) {
+      let found = false;
+      let confidence = "low";
+      if (check.windowKeys?.some((k) => k in window)) {
+        found = true;
+        confidence = "high";
+      }
+      if (!found && check.scriptPattern) {
+        const re = new RegExp(check.scriptPattern.source, "i");
+        const matchingSrc = scripts.find((s) => re.test(s));
+        if (matchingSrc) {
+          found = true;
+          confidence = "medium";
+        }
+      }
+      if (found) flagSdks.push({ sdk: check.sdk, confidence, details: void 0 });
+    }
+    const abTools = [];
+    if ("_vwo_code" in window || "_vis_opt_queue" in window) abTools.push({ tool: "VWO", experiments: [] });
+    if ("dataLayer" in window) {
+      const dl = window.dataLayer ?? [];
+      const expEvents = dl.filter((e) => e?.event?.includes("experiment") || e?.experimentId);
+      if (expEvents.length > 0) abTools.push({ tool: "Google Tag Manager Experiments", experiments: expEvents.slice(0, 5).map((e) => e.experimentId ?? "") });
+    }
+    return { flagSdks, abTools };
+  }, FLAG_SDK_CHECKS).catch(() => ({ flagSdks: [], abTools: [] }));
+  return {
+    url,
+    flagSdks: result.flagSdks,
+    abTools: result.abTools,
+    hasAnyFlags: result.flagSdks.length > 0 || result.abTools.length > 0,
+    detectedAt
+  };
+}
+export {
+  detectFeatureFlags
+};
+//# sourceMappingURL=feature-flag-detector.js.map

package/dist/feature-flag-detector.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/feature-flag-detector.ts"],
+  "sourcesContent": ["/**\n * Feature flag + A/B test detector.\n *\n * Detects popular feature flag / experiment SDKs loaded on the page:\n * LaunchDarkly, Optimizely, Split.io, Unleash, GrowthBook, Statsig, Flagsmith, Amplitude Experiment.\n *\n * Also detects common A/B test patterns: data-testid attributes, gtm.js experiment layers,\n * Optimizely variation IDs, VWO, etc.\n *\n * Why: ZeTa needs to know when a page shows different content to different users\n * so test generation can account for variants and not generate flaky tests.\n */\n\nimport type { Page } from 'playwright';\n\nexport interface FeatureFlagInfo {\n  sdk: string;\n  confidence: 'high' | 'medium' | 'low';\n  details?: string;\n}\n\nexport interface AbTestInfo {\n  tool: string;\n  experiments?: string[];\n}\n\nexport interface FeatureFlagDetectionResult {\n  url: string;\n  flagSdks: FeatureFlagInfo[];\n  abTools: AbTestInfo[];\n  hasAnyFlags: boolean;\n  detectedAt: string;\n}\n\n// SDK detectors \u2014 checked against window object + script URLs\nconst FLAG_SDK_CHECKS: Array<{\n  sdk: string;\n  windowKeys?: string[];\n  scriptPattern?: RegExp;\n}> = [\n  { sdk: 'LaunchDarkly', windowKeys: ['LDClient', 'LaunchDarkly', 'ldClient'], scriptPattern: /launchdarkly/i },\n  { sdk: 'Optimizely', windowKeys: ['optimizely', 'optimizelySdk'], scriptPattern: /optimizely/i },\n  { sdk: 'Split.io', windowKeys: ['splitio', 'SplitFactory'], scriptPattern: /split\\.io/i },\n  { sdk: 'GrowthBook', windowKeys: ['growthbook', 'GrowthBook'], scriptPattern: /growthbook/i },\n  { sdk: 'Statsig', windowKeys: ['statsig', 'StatsigClient'], scriptPattern: /statsig/i },\n  { sdk: 'Flagsmith', windowKeys: ['flagsmith'], scriptPattern: /flagsmith/i },\n  { sdk: 'Unleash', windowKeys: ['unleash', 'UnleashClient'], scriptPattern: /unleash/i },\n  { sdk: 'VWO', windowKeys: ['vwo', '_vwo'], scriptPattern: /vwo\\.com/i },\n  { sdk: 'Google Optimize', windowKeys: ['gaData', 'google_optimize'], scriptPattern: /googleoptimize/i },\n  { sdk: 'AB Tasty', windowKeys: ['ABTasty'], scriptPattern: /abtasty/i },\n];\n\nexport async function detectFeatureFlags(page: Page, url: string): Promise<FeatureFlagDetectionResult> {\n  const detectedAt = new Date().toISOString();\n\n  const result = await page.evaluate((checks) => {\n    const flagSdks: any[] = [];\n    const scripts = Array.from(document.querySelectorAll('script[src]')).map((s: any) => s.src ?? '');\n\n    for (const check of checks) {\n      let found = false;\n      let confidence = 'low';\n\n      // Check window keys\n      if (check.windowKeys?.some((k: string) => k in (window as any))) {\n        found = true;\n        confidence = 'high';\n      }\n      // Check script URLs\n      if (!found && check.scriptPattern) {\n        const re = new RegExp(check.scriptPattern.source, 'i');\n        const matchingSrc = scripts.find(s => re.test(s));\n        if (matchingSrc) { found = true; confidence = 'medium'; }\n      }\n      if (found) flagSdks.push({ sdk: check.sdk, confidence, details: undefined });\n    }\n\n    // A/B tools\n    const abTools: any[] = [];\n    if ('_vwo_code' in window || '_vis_opt_queue' in (window as any)) abTools.push({ tool: 'VWO', experiments: [] });\n    if ('dataLayer' in window) {\n      const dl = (window as any).dataLayer ?? [];\n      const expEvents = dl.filter((e: any) => e?.event?.includes('experiment') || e?.experimentId);\n      if (expEvents.length > 0) abTools.push({ tool: 'Google Tag Manager Experiments', experiments: expEvents.slice(0, 5).map((e: any) => e.experimentId ?? '') });\n    }\n\n    return { flagSdks, abTools };\n  }, FLAG_SDK_CHECKS as any).catch(() => ({ flagSdks: [], abTools: [] }));\n\n  return {\n    url,\n    flagSdks: result.flagSdks,\n    abTools: result.abTools,\n    hasAnyFlags: result.flagSdks.length > 0 || result.abTools.length > 0,\n    detectedAt,\n  };\n}\n"],
+  "mappings": "AAmCA,MAAM,kBAID;AAAA,EACH,EAAE,KAAK,gBAAgB,YAAY,CAAC,YAAY,gBAAgB,UAAU,GAAG,eAAe,gBAAgB;AAAA,EAC5G,EAAE,KAAK,cAAc,YAAY,CAAC,cAAc,eAAe,GAAG,eAAe,cAAc;AAAA,EAC/F,EAAE,KAAK,YAAY,YAAY,CAAC,WAAW,cAAc,GAAG,eAAe,aAAa;AAAA,EACxF,EAAE,KAAK,cAAc,YAAY,CAAC,cAAc,YAAY,GAAG,eAAe,cAAc;AAAA,EAC5F,EAAE,KAAK,WAAW,YAAY,CAAC,WAAW,eAAe,GAAG,eAAe,WAAW;AAAA,EACtF,EAAE,KAAK,aAAa,YAAY,CAAC,WAAW,GAAG,eAAe,aAAa;AAAA,EAC3E,EAAE,KAAK,WAAW,YAAY,CAAC,WAAW,eAAe,GAAG,eAAe,WAAW;AAAA,EACtF,EAAE,KAAK,OAAO,YAAY,CAAC,OAAO,MAAM,GAAG,eAAe,YAAY;AAAA,EACtE,EAAE,KAAK,mBAAmB,YAAY,CAAC,UAAU,iBAAiB,GAAG,eAAe,kBAAkB;AAAA,EACtG,EAAE,KAAK,YAAY,YAAY,CAAC,SAAS,GAAG,eAAe,WAAW;AACxE;AAEA,eAAsB,mBAAmB,MAAY,KAAkD;AACrG,QAAM,cAAa,oBAAI,KAAK,GAAE,YAAY;AAE1C,QAAM,SAAS,MAAM,KAAK,SAAS,CAAC,WAAW;AAC7C,UAAM,WAAkB,CAAC;AACzB,UAAM,UAAU,MAAM,KAAK,SAAS,iBAAiB,aAAa,CAAC,EAAE,IAAI,CAAC,MAAW,EAAE,OAAO,EAAE;AAEhG,eAAW,SAAS,QAAQ;AAC1B,UAAI,QAAQ;AACZ,UAAI,aAAa;AAGjB,UAAI,MAAM,YAAY,KAAK,CAAC,MAAc,KAAM,MAAc,GAAG;AAC/D,gBAAQ;AACR,qBAAa;AAAA,MACf;AAEA,UAAI,CAAC,SAAS,MAAM,eAAe;AACjC,cAAM,KAAK,IAAI,OAAO,MAAM,cAAc,QAAQ,GAAG;AACrD,cAAM,cAAc,QAAQ,KAAK,OAAK,GAAG,KAAK,CAAC,CAAC;AAChD,YAAI,aAAa;AAAE,kBAAQ;AAAM,uBAAa;AAAA,QAAU;AAAA,MAC1D;AACA,UAAI,MAAO,UAAS,KAAK,EAAE,KAAK,MAAM,KAAK,YAAY,SAAS,OAAU,CAAC;AAAA,IAC7E;AAGA,UAAM,UAAiB,CAAC;AACxB,QAAI,eAAe,UAAU,oBAAqB,OAAgB,SAAQ,KAAK,EAAE,MAAM,OAAO,aAAa,CAAC,EAAE,CAAC;AAC/G,QAAI,eAAe,QAAQ;AACzB,YAAM,KAAM,OAAe,aAAa,CAAC;AACzC,YAAM,YAAY,GAAG,OAAO,CAAC,MAAW,GAAG,OAAO,SAAS,YAAY,KAAK,GAAG,YAAY;AAC3F,UAAI,UAAU,SAAS,EAAG,SAAQ,KAAK,EAAE,MAAM,kCAAkC,aAAa,UAAU,MAAM,GAAG,CAAC,EAAE,IAAI,CAAC,MAAW,EAAE,gBAAgB,EAAE,EAAE,CAAC;AAAA,IAC7J;AAEA,WAAO,EAAE,UAAU,QAAQ;AAAA,EAC7B,GAAG,eAAsB,EAAE,MAAM,OAAO,EAAE,UAAU,CAAC,GAAG,SAAS,CAAC,EAAE,EAAE;AAEtE,SAAO;AAAA,IACL;AAAA,IACA,UAAU,OAAO;AAAA,IACjB,SAAS,OAAO;AAAA,IAChB,aAAa,OAAO,SAAS,SAAS,KAAK,OAAO,QAAQ,SAAS;AAAA,IACnE;AAAA,EACF;AACF;",
+  "names": []
+}

package/dist/filter-pagination-crawler.d.ts

@@ -0,0 +1,52 @@
+/**
+ * Filter state crawling and pagination capture.
+ *
+ * Discovers filter controls (select, checkbox-group, radio-group, search),
+ * applies each meaningful option and records result counts and first-row
+ * values, then detects and walks pagination (numbered, load-more, infinite
+ * scroll). Also captures sort column behaviour.
+ *
+ * All per-filter / per-page errors are caught and skipped so one bad control
+ * never aborts the full crawl.
+ */
+import type { Page } from 'playwright';
+export interface FilterOption {
+    value: string;
+    label: string;
+}
+export interface FilterDef {
+    kind: 'select' | 'checkbox-group' | 'radio-group' | 'search';
+    label: string;
+    locatorHint: string;
+    options: FilterOption[];
+}
+export interface FilterState {
+    filterLabel: string;
+    filterValue: string;
+    resultCount: number | null;
+    firstRowValues: string[];
+}
+export interface PaginationCapture {
+    kind: 'numbered' | 'load-more' | 'infinite-scroll' | 'none';
+    totalPages?: number;
+    totalItems?: number;
+    pageSizes?: number[];
+    pages: Array<{
+        pageNum: number;
+        rowValues: string[];
+    }>;
+}
+export interface FilterPaginationResult {
+    filters: FilterDef[];
+    filterStates: FilterState[];
+    pagination: PaginationCapture;
+    sortColumns: Array<{
+        header: string;
+        ariaSortAfterClick: string | null;
+    }>;
+    searchCapture?: {
+        noResultsText: string | null;
+        sampleSuggestions: string[];
+    };
+}
+export declare function crawlFiltersAndPagination(page: Page): Promise<FilterPaginationResult>;

package/dist/filter-pagination-crawler.js

@@ -0,0 +1,299 @@
+async function waitForSettle(page, ms = 1500) {
+  await Promise.race([
+    page.waitForLoadState("networkidle").catch(() => void 0),
+    new Promise((r) => setTimeout(r, ms))
+  ]);
+}
+async function captureFirstRow(page) {
+  try {
+    return await page.evaluate(() => {
+      const row = document.querySelector("tbody tr");
+      if (!row) return [];
+      return Array.from(row.querySelectorAll("td")).map((td) => td.innerText.trim());
+    });
+  } catch {
+    return [];
+  }
+}
+async function readResultCount(page) {
+  try {
+    const text = await page.evaluate(() => document.body.innerText);
+    const match = text.match(/(\d[\d,]*)\s*(result|item|entr|record)/i);
+    if (!match) return null;
+    return parseInt(match[1].replace(/,/g, ""), 10);
+  } catch {
+    return null;
+  }
+}
+async function discoverFilters(page) {
+  const filters = [];
+  try {
+    const selectHandles = await page.locator('select:not([role="dialog"] select)').all();
+    for (const handle of selectHandles.slice(0, 10)) {
+      try {
+        const id = await handle.getAttribute("id").catch(() => null);
+        const name = await handle.getAttribute("name").catch(() => null);
+        const locatorHint = id ? `#${id}` : name ? `select[name="${name}"]` : "select";
+        let label = "";
+        if (id) {
+          label = await page.locator(`label[for="${id}"]`).innerText().catch(() => "");
+        }
+        if (!label) {
+          label = await handle.evaluate((el) => {
+            const lbl = el.closest("label") ?? el.previousElementSibling;
+            return lbl?.innerText?.trim() ?? "";
+          }).catch(() => "");
+        }
+        if (!label) label = name ?? id ?? "filter";
+        const rawOptions = await handle.evaluate(
+          (el) => Array.from(el.options).map((o) => ({ value: o.value, label: o.text.trim() }))
+        ).catch(() => []);
+        const options = rawOptions.filter(
+          (o) => !(/^all$|^any$|^select|^--/i.test(o.label) || o.value === "")
+        );
+        if (options.length === 0) continue;
+        filters.push({ kind: "select", label, locatorHint, options });
+        if (filters.length >= 10) break;
+      } catch {
+      }
+    }
+  } catch {
+  }
+  if (filters.length < 10) {
+    try {
+      const searchHandles = await page.locator(
+        'input[type="search"]:not([role="dialog"] input), input[placeholder*="search" i]:not([role="dialog"] input)'
+      ).all();
+      for (const handle of searchHandles.slice(0, 10 - filters.length)) {
+        try {
+          const id = await handle.getAttribute("id").catch(() => null);
+          const placeholder = await handle.getAttribute("placeholder").catch(() => null);
+          const locatorHint = id ? `#${id}` : placeholder ? `input[placeholder="${placeholder}"]` : 'input[type="search"]';
+          let label = "";
+          if (id) {
+            label = await page.locator(`label[for="${id}"]`).innerText().catch(() => "");
+          }
+          if (!label) label = placeholder ?? "search";
+          filters.push({ kind: "search", label, locatorHint, options: [] });
+          if (filters.length >= 10) break;
+        } catch {
+        }
+      }
+    } catch {
+    }
+  }
+  return filters;
+}
+async function applyFilterAndCapture(page, filter, option) {
+  const state = {
+    filterLabel: filter.label,
+    filterValue: option.value,
+    resultCount: null,
+    firstRowValues: []
+  };
+  try {
+    if (filter.kind === "select") {
+      await page.locator(filter.locatorHint).selectOption(option.value);
+    } else if (filter.kind === "search") {
+      await page.locator(filter.locatorHint).fill(option.value);
+      await page.locator(filter.locatorHint).press("Enter");
+    }
+    await waitForSettle(page, 1500);
+    state.resultCount = await readResultCount(page);
+    state.firstRowValues = await captureFirstRow(page);
+  } catch {
+  }
+  try {
+    if (filter.kind === "select") {
+      const firstOption = await page.locator(filter.locatorHint).evaluate(
+        (el) => el.options[0]?.value ?? ""
+      );
+      await page.locator(filter.locatorHint).selectOption(firstOption);
+    } else if (filter.kind === "search") {
+      await page.locator(filter.locatorHint).fill("");
+      await page.locator(filter.locatorHint).press("Enter");
+    }
+    await waitForSettle(page, 800);
+  } catch {
+  }
+  return state;
+}
+async function crawlFilterStates(page, filters) {
+  const states = [];
+  let totalStates = 0;
+  for (const filter of filters) {
+    if (totalStates >= 15) break;
+    const { options } = filter;
+    if (options.length === 0) continue;
+    const indices = /* @__PURE__ */ new Set([
+      0,
+      Math.floor((options.length - 1) / 2),
+      options.length - 1
+    ]);
+    const picked = Array.from(indices).slice(0, 3);
+    for (const idx of picked) {
+      if (totalStates >= 15) break;
+      const option = options[idx];
+      try {
+        const s = await applyFilterAndCapture(page, filter, option);
+        states.push(s);
+      } catch {
+      }
+      totalStates++;
+    }
+  }
+  return states;
+}
+async function detectPaginationKind(page) {
+  try {
+    const numberedCount = await page.locator('[aria-label*="page" i], .pagination button, [role="navigation"] button').count().catch(() => 0);
+    if (numberedCount > 1) return "numbered";
+    const loadMore = await page.locator('button:has-text("Load more"), button:has-text("Show more"), a:has-text("Load more")').count().catch(() => 0);
+    if (loadMore > 0) return "load-more";
+    const isInfinite = await page.evaluate(() => {
+      return document.documentElement.scrollHeight > window.innerHeight * 2;
+    }).catch(() => false);
+    if (isInfinite) return "infinite-scroll";
+  } catch {
+  }
+  return "none";
+}
+async function crawlNumberedPagination(page) {
+  const pages = [];
+  const pageSizes = [];
+  try {
+    const sizeSelect = page.locator('select[name*="size" i], select[name*="per" i], select[aria-label*="per page" i]');
+    const sizeCount = await sizeSelect.count();
+    if (sizeCount > 0) {
+      const rawSizes = await sizeSelect.first().evaluate(
+        (el) => Array.from(el.options).map((o) => parseInt(o.value, 10)).filter((n) => !isNaN(n))
+      );
+      pageSizes.push(...rawSizes);
+    }
+  } catch {
+  }
+  let totalPages;
+  try {
+    totalPages = await page.evaluate(() => {
+      const buttons = Array.from(
+        document.querySelectorAll('[aria-label*="page" i], .pagination button')
+      );
+      const nums = buttons.map((b) => parseInt(b.innerText.trim(), 10)).filter((n) => !isNaN(n));
+      return nums.length > 0 ? Math.max(...nums) : void 0;
+    });
+  } catch {
+  }
+  pages.push({ pageNum: 1, rowValues: await captureFirstRow(page) });
+  for (let pageNum = 2; pageNum <= 20; pageNum++) {
+    try {
+      const nextBtn = page.locator(
+        'button[aria-label*="next" i], a[aria-label*="next" i], [role="navigation"] button:last-child'
+      ).first();
+      const isDisabled = await nextBtn.isDisabled().catch(() => true);
+      if (isDisabled) break;
+      await nextBtn.click();
+      await waitForSettle(page, 2e3);
+      const rowValues = await captureFirstRow(page);
+      pages.push({ pageNum, rowValues });
+    } catch {
+      break;
+    }
+  }
+  return {
+    kind: "numbered",
+    totalPages,
+    pageSizes: pageSizes.length > 0 ? pageSizes : void 0,
+    pages
+  };
+}
+async function captureInfiniteScroll(page) {
+  const pages = [];
+  let prevRowCount = -1;
+  let iteration = 0;
+  while (iteration < 10) {
+    try {
+      const rowCount = await page.evaluate(
+        () => document.querySelectorAll("tbody tr").length
+      ).catch(() => 0);
+      if (rowCount === prevRowCount) break;
+      prevRowCount = rowCount;
+      pages.push({ pageNum: iteration + 1, rowValues: await captureFirstRow(page) });
+      await page.evaluate(() => window.scrollTo(0, document.body.scrollHeight));
+      await new Promise((r) => setTimeout(r, 800));
+    } catch {
+      break;
+    }
+    iteration++;
+  }
+  const totalItems = await page.evaluate(
+    () => document.querySelectorAll("tbody tr").length
+  ).catch(() => void 0);
+  return {
+    kind: "infinite-scroll",
+    totalItems: typeof totalItems === "number" ? totalItems : void 0,
+    pages
+  };
+}
+async function captureSortColumns(page) {
+  const results = [];
+  try {
+    const sortable = await page.locator("[aria-sort]").all();
+    for (const th of sortable.slice(0, 5)) {
+      try {
+        const header = await th.innerText().catch(() => "");
+        await th.click();
+        await waitForSettle(page, 800);
+        const ariaSortAfterClick = await th.getAttribute("aria-sort").catch(() => null);
+        results.push({ header: header.trim(), ariaSortAfterClick });
+        await th.click().catch(() => void 0);
+        await waitForSettle(page, 600);
+      } catch {
+      }
+    }
+  } catch {
+  }
+  return results;
+}
+async function crawlFiltersAndPagination(page) {
+  let filters = [];
+  try {
+    filters = await discoverFilters(page);
+  } catch {
+  }
+  let filterStates = [];
+  try {
+    filterStates = await crawlFilterStates(page, filters);
+  } catch {
+  }
+  let pagination = { kind: "none", pages: [] };
+  try {
+    const kind = await detectPaginationKind(page);
+    if (kind === "numbered") {
+      pagination = await crawlNumberedPagination(page);
+    } else if (kind === "load-more") {
+      pagination = await captureInfiniteScroll(page);
+      pagination = { ...pagination, kind: "load-more" };
+    } else if (kind === "infinite-scroll") {
+      pagination = await captureInfiniteScroll(page);
+    } else {
+      const rowValues = await captureFirstRow(page);
+      pagination = { kind: "none", pages: rowValues.length > 0 ? [{ pageNum: 1, rowValues }] : [] };
+    }
+  } catch {
+  }
+  let sortColumns = [];
+  try {
+    sortColumns = await captureSortColumns(page);
+  } catch {
+  }
+  return {
+    filters,
+    filterStates,
+    pagination,
+    sortColumns
+  };
+}
+export {
+  crawlFiltersAndPagination
+};
+//# sourceMappingURL=filter-pagination-crawler.js.map

package/dist/filter-pagination-crawler.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/filter-pagination-crawler.ts"],
+  "sourcesContent": ["/**\n * Filter state crawling and pagination capture.\n *\n * Discovers filter controls (select, checkbox-group, radio-group, search),\n * applies each meaningful option and records result counts and first-row\n * values, then detects and walks pagination (numbered, load-more, infinite\n * scroll). Also captures sort column behaviour.\n *\n * All per-filter / per-page errors are caught and skipped so one bad control\n * never aborts the full crawl.\n */\n\nimport type { Page } from 'playwright';\n\n// ---------------------------------------------------------------------------\n// Public interfaces\n// ---------------------------------------------------------------------------\n\nexport interface FilterOption {\n  value: string;\n  label: string;\n}\n\nexport interface FilterDef {\n  kind: 'select' | 'checkbox-group' | 'radio-group' | 'search';\n  label: string;\n  locatorHint: string;\n  options: FilterOption[];\n}\n\nexport interface FilterState {\n  filterLabel: string;\n  filterValue: string;\n  resultCount: number | null;\n  firstRowValues: string[];\n}\n\nexport interface PaginationCapture {\n  kind: 'numbered' | 'load-more' | 'infinite-scroll' | 'none';\n  totalPages?: number;\n  totalItems?: number;\n  pageSizes?: number[];\n  pages: Array<{ pageNum: number; rowValues: string[] }>;\n}\n\nexport interface FilterPaginationResult {\n  filters: FilterDef[];\n  filterStates: FilterState[];\n  pagination: PaginationCapture;\n  sortColumns: Array<{ header: string; ariaSortAfterClick: string | null }>;\n  searchCapture?: { noResultsText: string | null; sampleSuggestions: string[] };\n}\n\n// ---------------------------------------------------------------------------\n// Internal helpers\n// ---------------------------------------------------------------------------\n\n/** Pause for up to `ms` ms or until network is idle, whichever comes first. */\nasync function waitForSettle(page: Page, ms = 1500): Promise<void> {\n  await Promise.race([\n    page.waitForLoadState('networkidle').catch(() => undefined),\n    new Promise<void>(r => setTimeout(r, ms)),\n  ]);\n}\n\n/** Extract visible cell text from the first tbody row, if any. */\nasync function captureFirstRow(page: Page): Promise<string[]> {\n  try {\n    return await page.evaluate(() => {\n      const row = document.querySelector('tbody tr');\n      if (!row) return [];\n      return Array.from(row.querySelectorAll('td')).map(td => td.innerText.trim());\n    });\n  } catch {\n    return [];\n  }\n}\n\n/** Read a result-count hint from visible page text using common patterns. */\nasync function readResultCount(page: Page): Promise<number | null> {\n  try {\n    const text = await page.evaluate(() => document.body.innerText);\n    const match = text.match(/(\\d[\\d,]*)\\s*(result|item|entr|record)/i);\n    if (!match) return null;\n    return parseInt(match[1].replace(/,/g, ''), 10);\n  } catch {\n    return null;\n  }\n}\n\n// ---------------------------------------------------------------------------\n// discoverFilters\n// ---------------------------------------------------------------------------\n\nasync function discoverFilters(page: Page): Promise<FilterDef[]> {\n  const filters: FilterDef[] = [];\n\n  // --- native <select> elements outside dialogs ---------------------------------\n  try {\n    const selectHandles = await page.locator('select:not([role=\"dialog\"] select)').all();\n    for (const handle of selectHandles.slice(0, 10)) {\n      try {\n        const id = await handle.getAttribute('id').catch(() => null);\n        const name = await handle.getAttribute('name').catch(() => null);\n        const locatorHint = id ? `#${id}` : name ? `select[name=\"${name}\"]` : 'select';\n\n        // Find associated label\n        let label = '';\n        if (id) {\n          label = await page.locator(`label[for=\"${id}\"]`).innerText().catch(() => '');\n        }\n        if (!label) {\n          label = await handle.evaluate((el: HTMLSelectElement) => {\n            const lbl = el.closest('label') ?? el.previousElementSibling;\n            return (lbl as HTMLElement | null)?.innerText?.trim() ?? '';\n          }).catch(() => '');\n        }\n        if (!label) label = name ?? id ?? 'filter';\n\n        // Collect options, skipping default/placeholder entries\n        const rawOptions: FilterOption[] = await handle.evaluate((el: HTMLSelectElement) =>\n          Array.from(el.options).map(o => ({ value: o.value, label: o.text.trim() })),\n        ).catch(() => []);\n\n        const options = rawOptions.filter(\n          o => !(/^all$|^any$|^select|^--/i.test(o.label) || o.value === ''),\n        );\n\n        if (options.length === 0) continue;\n\n        filters.push({ kind: 'select', label, locatorHint, options });\n        if (filters.length >= 10) break;\n      } catch {\n        // skip this select\n      }\n    }\n  } catch {\n    // page.locator failed\n  }\n\n  // --- search inputs ------------------------------------------------------------\n  if (filters.length < 10) {\n    try {\n      const searchHandles = await page\n        .locator(\n          'input[type=\"search\"]:not([role=\"dialog\"] input), ' +\n          'input[placeholder*=\"search\" i]:not([role=\"dialog\"] input)',\n        )\n        .all();\n\n      for (const handle of searchHandles.slice(0, 10 - filters.length)) {\n        try {\n          const id = await handle.getAttribute('id').catch(() => null);\n          const placeholder = await handle.getAttribute('placeholder').catch(() => null);\n          const locatorHint = id ? `#${id}` : placeholder ? `input[placeholder=\"${placeholder}\"]` : 'input[type=\"search\"]';\n\n          let label = '';\n          if (id) {\n            label = await page.locator(`label[for=\"${id}\"]`).innerText().catch(() => '');\n          }\n          if (!label) label = placeholder ?? 'search';\n\n          filters.push({ kind: 'search', label, locatorHint, options: [] });\n          if (filters.length >= 10) break;\n        } catch {\n          // skip\n        }\n      }\n    } catch {\n      // locator failed\n    }\n  }\n\n  return filters;\n}\n\n// ---------------------------------------------------------------------------\n// applyFilterAndCapture\n// ---------------------------------------------------------------------------\n\nasync function applyFilterAndCapture(\n  page: Page,\n  filter: FilterDef,\n  option: FilterOption,\n): Promise<FilterState> {\n  const state: FilterState = {\n    filterLabel: filter.label,\n    filterValue: option.value,\n    resultCount: null,\n    firstRowValues: [],\n  };\n\n  try {\n    if (filter.kind === 'select') {\n      await page.locator(filter.locatorHint).selectOption(option.value);\n    } else if (filter.kind === 'search') {\n      await page.locator(filter.locatorHint).fill(option.value);\n      await page.locator(filter.locatorHint).press('Enter');\n    }\n\n    await waitForSettle(page, 1500);\n\n    state.resultCount = await readResultCount(page);\n    state.firstRowValues = await captureFirstRow(page);\n  } catch {\n    // non-fatal: return partial state\n  }\n\n  // Reset filter\n  try {\n    if (filter.kind === 'select') {\n      const firstOption = await page.locator(filter.locatorHint).evaluate(\n        (el: HTMLSelectElement) => el.options[0]?.value ?? '',\n      );\n      await page.locator(filter.locatorHint).selectOption(firstOption);\n    } else if (filter.kind === 'search') {\n      await page.locator(filter.locatorHint).fill('');\n      await page.locator(filter.locatorHint).press('Enter');\n    }\n    await waitForSettle(page, 800);\n  } catch {\n    // ignore reset errors\n  }\n\n  return state;\n}\n\n// ---------------------------------------------------------------------------\n// crawlFilterStates\n// ---------------------------------------------------------------------------\n\nasync function crawlFilterStates(page: Page, filters: FilterDef[]): Promise<FilterState[]> {\n  const states: FilterState[] = [];\n  let totalStates = 0;\n\n  for (const filter of filters) {\n    if (totalStates >= 15) break;\n\n    const { options } = filter;\n    if (options.length === 0) continue;\n\n    // Pick first, middle, last \u2014 deduplicated, max 3\n    const indices = new Set<number>([\n      0,\n      Math.floor((options.length - 1) / 2),\n      options.length - 1,\n    ]);\n    const picked = Array.from(indices).slice(0, 3);\n\n    for (const idx of picked) {\n      if (totalStates >= 15) break;\n      const option = options[idx];\n      try {\n        const s = await applyFilterAndCapture(page, filter, option);\n        states.push(s);\n      } catch {\n        // skip\n      }\n      totalStates++;\n    }\n  }\n\n  return states;\n}\n\n// ---------------------------------------------------------------------------\n// detectPaginationKind\n// ---------------------------------------------------------------------------\n\nasync function detectPaginationKind(\n  page: Page,\n): Promise<'numbered' | 'load-more' | 'infinite-scroll' | 'none'> {\n  try {\n    // Numbered pagination: ARIA page buttons or .pagination\n    const numberedCount = await page\n      .locator('[aria-label*=\"page\" i], .pagination button, [role=\"navigation\"] button')\n      .count()\n      .catch(() => 0);\n    if (numberedCount > 1) return 'numbered';\n\n    // Load-more button\n    const loadMore = await page\n      .locator('button:has-text(\"Load more\"), button:has-text(\"Show more\"), a:has-text(\"Load more\")')\n      .count()\n      .catch(() => 0);\n    if (loadMore > 0) return 'load-more';\n\n    // Infinite scroll: scrollable content significantly taller than viewport\n    const isInfinite = await page.evaluate(() => {\n      return document.documentElement.scrollHeight > window.innerHeight * 2;\n    }).catch(() => false);\n    if (isInfinite) return 'infinite-scroll';\n  } catch {\n    // fall through\n  }\n\n  return 'none';\n}\n\n// ---------------------------------------------------------------------------\n// crawlNumberedPagination\n// ---------------------------------------------------------------------------\n\nasync function crawlNumberedPagination(page: Page): Promise<PaginationCapture> {\n  const pages: Array<{ pageNum: number; rowValues: string[] }> = [];\n  const pageSizes: number[] = [];\n\n  // Read page size selector if present\n  try {\n    const sizeSelect = page.locator('select[name*=\"size\" i], select[name*=\"per\" i], select[aria-label*=\"per page\" i]');\n    const sizeCount = await sizeSelect.count();\n    if (sizeCount > 0) {\n      const rawSizes = await sizeSelect.first().evaluate((el: HTMLSelectElement) =>\n        Array.from(el.options).map(o => parseInt(o.value, 10)).filter(n => !isNaN(n)),\n      );\n      pageSizes.push(...rawSizes);\n    }\n  } catch {\n    // ignore\n  }\n\n  // Determine total pages from pagination widget\n  let totalPages: number | undefined;\n  try {\n    totalPages = await page.evaluate(() => {\n      const buttons = Array.from(\n        document.querySelectorAll('[aria-label*=\"page\" i], .pagination button'),\n      );\n      const nums = buttons\n        .map(b => parseInt((b as HTMLElement).innerText.trim(), 10))\n        .filter(n => !isNaN(n));\n      return nums.length > 0 ? Math.max(...nums) : undefined;\n    });\n  } catch {\n    // leave undefined\n  }\n\n  // Capture first page immediately\n  pages.push({ pageNum: 1, rowValues: await captureFirstRow(page) });\n\n  // Walk pages by clicking Next\n  for (let pageNum = 2; pageNum <= 20; pageNum++) {\n    try {\n      const nextBtn = page.locator(\n        'button[aria-label*=\"next\" i], a[aria-label*=\"next\" i], [role=\"navigation\"] button:last-child',\n      ).first();\n\n      const isDisabled = await nextBtn.isDisabled().catch(() => true);\n      if (isDisabled) break;\n\n      await nextBtn.click();\n      await waitForSettle(page, 2000);\n\n      const rowValues = await captureFirstRow(page);\n      pages.push({ pageNum, rowValues });\n    } catch {\n      break;\n    }\n  }\n\n  return {\n    kind: 'numbered',\n    totalPages,\n    pageSizes: pageSizes.length > 0 ? pageSizes : undefined,\n    pages,\n  };\n}\n\n// ---------------------------------------------------------------------------\n// captureInfiniteScroll\n// ---------------------------------------------------------------------------\n\nasync function captureInfiniteScroll(page: Page): Promise<PaginationCapture> {\n  const pages: Array<{ pageNum: number; rowValues: string[] }> = [];\n\n  let prevRowCount = -1;\n  let iteration = 0;\n\n  while (iteration < 10) {\n    try {\n      const rowCount = await page.evaluate(() =>\n        document.querySelectorAll('tbody tr').length,\n      ).catch(() => 0);\n\n      if (rowCount === prevRowCount) break; // stable\n      prevRowCount = rowCount;\n\n      // Record current first-row snapshot per scroll step\n      pages.push({ pageNum: iteration + 1, rowValues: await captureFirstRow(page) });\n\n      await page.evaluate(() => window.scrollTo(0, document.body.scrollHeight));\n      await new Promise<void>(r => setTimeout(r, 800));\n    } catch {\n      break;\n    }\n    iteration++;\n  }\n\n  const totalItems = await page.evaluate(() =>\n    document.querySelectorAll('tbody tr').length,\n  ).catch(() => undefined);\n\n  return {\n    kind: 'infinite-scroll',\n    totalItems: typeof totalItems === 'number' ? totalItems : undefined,\n    pages,\n  };\n}\n\n// ---------------------------------------------------------------------------\n// captureSortColumns\n// ---------------------------------------------------------------------------\n\nasync function captureSortColumns(\n  page: Page,\n): Promise<Array<{ header: string; ariaSortAfterClick: string | null }>> {\n  const results: Array<{ header: string; ariaSortAfterClick: string | null }> = [];\n\n  try {\n    const sortable = await page.locator('[aria-sort]').all();\n    for (const th of sortable.slice(0, 5)) {\n      try {\n        const header = await th.innerText().catch(() => '');\n        await th.click();\n        await waitForSettle(page, 800);\n        const ariaSortAfterClick = await th.getAttribute('aria-sort').catch(() => null);\n        results.push({ header: header.trim(), ariaSortAfterClick });\n\n        // Reset by clicking again (most tables toggle back)\n        await th.click().catch(() => undefined);\n        await waitForSettle(page, 600);\n      } catch {\n        // skip this column\n      }\n    }\n  } catch {\n    // no sortable columns found\n  }\n\n  return results;\n}\n\n// ---------------------------------------------------------------------------\n// crawlFiltersAndPagination \u2014 public orchestrator\n// ---------------------------------------------------------------------------\n\nexport async function crawlFiltersAndPagination(page: Page): Promise<FilterPaginationResult> {\n  // 1. Discover filters\n  let filters: FilterDef[] = [];\n  try {\n    filters = await discoverFilters(page);\n  } catch {\n    // proceed with empty filters\n  }\n\n  // 2. Apply filters and capture states\n  let filterStates: FilterState[] = [];\n  try {\n    filterStates = await crawlFilterStates(page, filters);\n  } catch {\n    // proceed with empty states\n  }\n\n  // 3. Detect and crawl pagination\n  let pagination: PaginationCapture = { kind: 'none', pages: [] };\n  try {\n    const kind = await detectPaginationKind(page);\n\n    if (kind === 'numbered') {\n      pagination = await crawlNumberedPagination(page);\n    } else if (kind === 'load-more') {\n      // Treat load-more similarly to infinite scroll capture\n      pagination = await captureInfiniteScroll(page);\n      pagination = { ...pagination, kind: 'load-more' };\n    } else if (kind === 'infinite-scroll') {\n      pagination = await captureInfiniteScroll(page);\n    } else {\n      // Capture single-page rows\n      const rowValues = await captureFirstRow(page);\n      pagination = { kind: 'none', pages: rowValues.length > 0 ? [{ pageNum: 1, rowValues }] : [] };\n    }\n  } catch {\n    // leave pagination as none\n  }\n\n  // 4. Sort columns\n  let sortColumns: Array<{ header: string; ariaSortAfterClick: string | null }> = [];\n  try {\n    sortColumns = await captureSortColumns(page);\n  } catch {\n    // proceed with empty sort\n  }\n\n  return {\n    filters,\n    filterStates,\n    pagination,\n    sortColumns,\n  };\n}\n"],
+  "mappings": "AA0DA,eAAe,cAAc,MAAY,KAAK,MAAqB;AACjE,QAAM,QAAQ,KAAK;AAAA,IACjB,KAAK,iBAAiB,aAAa,EAAE,MAAM,MAAM,MAAS;AAAA,IAC1D,IAAI,QAAc,OAAK,WAAW,GAAG,EAAE,CAAC;AAAA,EAC1C,CAAC;AACH;AAGA,eAAe,gBAAgB,MAA+B;AAC5D,MAAI;AACF,WAAO,MAAM,KAAK,SAAS,MAAM;AAC/B,YAAM,MAAM,SAAS,cAAc,UAAU;AAC7C,UAAI,CAAC,IAAK,QAAO,CAAC;AAClB,aAAO,MAAM,KAAK,IAAI,iBAAiB,IAAI,CAAC,EAAE,IAAI,QAAM,GAAG,UAAU,KAAK,CAAC;AAAA,IAC7E,CAAC;AAAA,EACH,QAAQ;AACN,WAAO,CAAC;AAAA,EACV;AACF;AAGA,eAAe,gBAAgB,MAAoC;AACjE,MAAI;AACF,UAAM,OAAO,MAAM,KAAK,SAAS,MAAM,SAAS,KAAK,SAAS;AAC9D,UAAM,QAAQ,KAAK,MAAM,yCAAyC;AAClE,QAAI,CAAC,MAAO,QAAO;AACnB,WAAO,SAAS,MAAM,CAAC,EAAE,QAAQ,MAAM,EAAE,GAAG,EAAE;AAAA,EAChD,QAAQ;AACN,WAAO;AAAA,EACT;AACF;AAMA,eAAe,gBAAgB,MAAkC;AAC/D,QAAM,UAAuB,CAAC;AAG9B,MAAI;AACF,UAAM,gBAAgB,MAAM,KAAK,QAAQ,oCAAoC,EAAE,IAAI;AACnF,eAAW,UAAU,cAAc,MAAM,GAAG,EAAE,GAAG;AAC/C,UAAI;AACF,cAAM,KAAK,MAAM,OAAO,aAAa,IAAI,EAAE,MAAM,MAAM,IAAI;AAC3D,cAAM,OAAO,MAAM,OAAO,aAAa,MAAM,EAAE,MAAM,MAAM,IAAI;AAC/D,cAAM,cAAc,KAAK,IAAI,EAAE,KAAK,OAAO,gBAAgB,IAAI,OAAO;AAGtE,YAAI,QAAQ;AACZ,YAAI,IAAI;AACN,kBAAQ,MAAM,KAAK,QAAQ,cAAc,EAAE,IAAI,EAAE,UAAU,EAAE,MAAM,MAAM,EAAE;AAAA,QAC7E;AACA,YAAI,CAAC,OAAO;AACV,kBAAQ,MAAM,OAAO,SAAS,CAAC,OAA0B;AACvD,kBAAM,MAAM,GAAG,QAAQ,OAAO,KAAK,GAAG;AACtC,mBAAQ,KAA4B,WAAW,KAAK,KAAK;AAAA,UAC3D,CAAC,EAAE,MAAM,MAAM,EAAE;AAAA,QACnB;AACA,YAAI,CAAC,MAAO,SAAQ,QAAQ,MAAM;AAGlC,cAAM,aAA6B,MAAM,OAAO;AAAA,UAAS,CAAC,OACxD,MAAM,KAAK,GAAG,OAAO,EAAE,IAAI,QAAM,EAAE,OAAO,EAAE,OAAO,OAAO,EAAE,KAAK,KAAK,EAAE,EAAE;AAAA,QAC5E,EAAE,MAAM,MAAM,CAAC,CAAC;AAEhB,cAAM,UAAU,WAAW;AAAA,UACzB,OAAK,EAAE,2BAA2B,KAAK,EAAE,KAAK,KAAK,EAAE,UAAU;AAAA,QACjE;AAEA,YAAI,QAAQ,WAAW,EAAG;AAE1B,gBAAQ,KAAK,EAAE,MAAM,UAAU,OAAO,aAAa,QAAQ,CAAC;AAC5D,YAAI,QAAQ,UAAU,GAAI;AAAA,MAC5B,QAAQ;AAAA,MAER;AAAA,IACF;AAAA,EACF,QAAQ;AAAA,EAER;AAGA,MAAI,QAAQ,SAAS,IAAI;AACvB,QAAI;AACF,YAAM,gBAAgB,MAAM,KACzB;AAAA,QACC;AAAA,MAEF,EACC,IAAI;AAEP,iBAAW,UAAU,cAAc,MAAM,GAAG,KAAK,QAAQ,MAAM,GAAG;AAChE,YAAI;AACF,gBAAM,KAAK,MAAM,OAAO,aAAa,IAAI,EAAE,MAAM,MAAM,IAAI;AAC3D,gBAAM,cAAc,MAAM,OAAO,aAAa,aAAa,EAAE,MAAM,MAAM,IAAI;AAC7E,gBAAM,cAAc,KAAK,IAAI,EAAE,KAAK,cAAc,sBAAsB,WAAW,OAAO;AAE1F,cAAI,QAAQ;AACZ,cAAI,IAAI;AACN,oBAAQ,MAAM,KAAK,QAAQ,cAAc,EAAE,IAAI,EAAE,UAAU,EAAE,MAAM,MAAM,EAAE;AAAA,UAC7E;AACA,cAAI,CAAC,MAAO,SAAQ,eAAe;AAEnC,kBAAQ,KAAK,EAAE,MAAM,UAAU,OAAO,aAAa,SAAS,CAAC,EAAE,CAAC;AAChE,cAAI,QAAQ,UAAU,GAAI;AAAA,QAC5B,QAAQ;AAAA,QAER;AAAA,MACF;AAAA,IACF,QAAQ;AAAA,IAER;AAAA,EACF;AAEA,SAAO;AACT;AAMA,eAAe,sBACb,MACA,QACA,QACsB;AACtB,QAAM,QAAqB;AAAA,IACzB,aAAa,OAAO;AAAA,IACpB,aAAa,OAAO;AAAA,IACpB,aAAa;AAAA,IACb,gBAAgB,CAAC;AAAA,EACnB;AAEA,MAAI;AACF,QAAI,OAAO,SAAS,UAAU;AAC5B,YAAM,KAAK,QAAQ,OAAO,WAAW,EAAE,aAAa,OAAO,KAAK;AAAA,IAClE,WAAW,OAAO,SAAS,UAAU;AACnC,YAAM,KAAK,QAAQ,OAAO,WAAW,EAAE,KAAK,OAAO,KAAK;AACxD,YAAM,KAAK,QAAQ,OAAO,WAAW,EAAE,MAAM,OAAO;AAAA,IACtD;AAEA,UAAM,cAAc,MAAM,IAAI;AAE9B,UAAM,cAAc,MAAM,gBAAgB,IAAI;AAC9C,UAAM,iBAAiB,MAAM,gBAAgB,IAAI;AAAA,EACnD,QAAQ;AAAA,EAER;AAGA,MAAI;AACF,QAAI,OAAO,SAAS,UAAU;AAC5B,YAAM,cAAc,MAAM,KAAK,QAAQ,OAAO,WAAW,EAAE;AAAA,QACzD,CAAC,OAA0B,GAAG,QAAQ,CAAC,GAAG,SAAS;AAAA,MACrD;AACA,YAAM,KAAK,QAAQ,OAAO,WAAW,EAAE,aAAa,WAAW;AAAA,IACjE,WAAW,OAAO,SAAS,UAAU;AACnC,YAAM,KAAK,QAAQ,OAAO,WAAW,EAAE,KAAK,EAAE;AAC9C,YAAM,KAAK,QAAQ,OAAO,WAAW,EAAE,MAAM,OAAO;AAAA,IACtD;AACA,UAAM,cAAc,MAAM,GAAG;AAAA,EAC/B,QAAQ;AAAA,EAER;AAEA,SAAO;AACT;AAMA,eAAe,kBAAkB,MAAY,SAA8C;AACzF,QAAM,SAAwB,CAAC;AAC/B,MAAI,cAAc;AAElB,aAAW,UAAU,SAAS;AAC5B,QAAI,eAAe,GAAI;AAEvB,UAAM,EAAE,QAAQ,IAAI;AACpB,QAAI,QAAQ,WAAW,EAAG;AAG1B,UAAM,UAAU,oBAAI,IAAY;AAAA,MAC9B;AAAA,MACA,KAAK,OAAO,QAAQ,SAAS,KAAK,CAAC;AAAA,MACnC,QAAQ,SAAS;AAAA,IACnB,CAAC;AACD,UAAM,SAAS,MAAM,KAAK,OAAO,EAAE,MAAM,GAAG,CAAC;AAE7C,eAAW,OAAO,QAAQ;AACxB,UAAI,eAAe,GAAI;AACvB,YAAM,SAAS,QAAQ,GAAG;AAC1B,UAAI;AACF,cAAM,IAAI,MAAM,sBAAsB,MAAM,QAAQ,MAAM;AAC1D,eAAO,KAAK,CAAC;AAAA,MACf,QAAQ;AAAA,MAER;AACA;AAAA,IACF;AAAA,EACF;AAEA,SAAO;AACT;AAMA,eAAe,qBACb,MACgE;AAChE,MAAI;AAEF,UAAM,gBAAgB,MAAM,KACzB,QAAQ,wEAAwE,EAChF,MAAM,EACN,MAAM,MAAM,CAAC;AAChB,QAAI,gBAAgB,EAAG,QAAO;AAG9B,UAAM,WAAW,MAAM,KACpB,QAAQ,qFAAqF,EAC7F,MAAM,EACN,MAAM,MAAM,CAAC;AAChB,QAAI,WAAW,EAAG,QAAO;AAGzB,UAAM,aAAa,MAAM,KAAK,SAAS,MAAM;AAC3C,aAAO,SAAS,gBAAgB,eAAe,OAAO,cAAc;AAAA,IACtE,CAAC,EAAE,MAAM,MAAM,KAAK;AACpB,QAAI,WAAY,QAAO;AAAA,EACzB,QAAQ;AAAA,EAER;AAEA,SAAO;AACT;AAMA,eAAe,wBAAwB,MAAwC;AAC7E,QAAM,QAAyD,CAAC;AAChE,QAAM,YAAsB,CAAC;AAG7B,MAAI;AACF,UAAM,aAAa,KAAK,QAAQ,iFAAiF;AACjH,UAAM,YAAY,MAAM,WAAW,MAAM;AACzC,QAAI,YAAY,GAAG;AACjB,YAAM,WAAW,MAAM,WAAW,MAAM,EAAE;AAAA,QAAS,CAAC,OAClD,MAAM,KAAK,GAAG,OAAO,EAAE,IAAI,OAAK,SAAS,EAAE,OAAO,EAAE,CAAC,EAAE,OAAO,OAAK,CAAC,MAAM,CAAC,CAAC;AAAA,MAC9E;AACA,gBAAU,KAAK,GAAG,QAAQ;AAAA,IAC5B;AAAA,EACF,QAAQ;AAAA,EAER;AAGA,MAAI;AACJ,MAAI;AACF,iBAAa,MAAM,KAAK,SAAS,MAAM;AACrC,YAAM,UAAU,MAAM;AAAA,QACpB,SAAS,iBAAiB,4CAA4C;AAAA,MACxE;AACA,YAAM,OAAO,QACV,IAAI,OAAK,SAAU,EAAkB,UAAU,KAAK,GAAG,EAAE,CAAC,EAC1D,OAAO,OAAK,CAAC,MAAM,CAAC,CAAC;AACxB,aAAO,KAAK,SAAS,IAAI,KAAK,IAAI,GAAG,IAAI,IAAI;AAAA,IAC/C,CAAC;AAAA,EACH,QAAQ;AAAA,EAER;AAGA,QAAM,KAAK,EAAE,SAAS,GAAG,WAAW,MAAM,gBAAgB,IAAI,EAAE,CAAC;AAGjE,WAAS,UAAU,GAAG,WAAW,IAAI,WAAW;AAC9C,QAAI;AACF,YAAM,UAAU,KAAK;AAAA,QACnB;AAAA,MACF,EAAE,MAAM;AAER,YAAM,aAAa,MAAM,QAAQ,WAAW,EAAE,MAAM,MAAM,IAAI;AAC9D,UAAI,WAAY;AAEhB,YAAM,QAAQ,MAAM;AACpB,YAAM,cAAc,MAAM,GAAI;AAE9B,YAAM,YAAY,MAAM,gBAAgB,IAAI;AAC5C,YAAM,KAAK,EAAE,SAAS,UAAU,CAAC;AAAA,IACnC,QAAQ;AACN;AAAA,IACF;AAAA,EACF;AAEA,SAAO;AAAA,IACL,MAAM;AAAA,IACN;AAAA,IACA,WAAW,UAAU,SAAS,IAAI,YAAY;AAAA,IAC9C;AAAA,EACF;AACF;AAMA,eAAe,sBAAsB,MAAwC;AAC3E,QAAM,QAAyD,CAAC;AAEhE,MAAI,eAAe;AACnB,MAAI,YAAY;AAEhB,SAAO,YAAY,IAAI;AACrB,QAAI;AACF,YAAM,WAAW,MAAM,KAAK;AAAA,QAAS,MACnC,SAAS,iBAAiB,UAAU,EAAE;AAAA,MACxC,EAAE,MAAM,MAAM,CAAC;AAEf,UAAI,aAAa,aAAc;AAC/B,qBAAe;AAGf,YAAM,KAAK,EAAE,SAAS,YAAY,GAAG,WAAW,MAAM,gBAAgB,IAAI,EAAE,CAAC;AAE7E,YAAM,KAAK,SAAS,MAAM,OAAO,SAAS,GAAG,SAAS,KAAK,YAAY,CAAC;AACxE,YAAM,IAAI,QAAc,OAAK,WAAW,GAAG,GAAG,CAAC;AAAA,IACjD,QAAQ;AACN;AAAA,IACF;AACA;AAAA,EACF;AAEA,QAAM,aAAa,MAAM,KAAK;AAAA,IAAS,MACrC,SAAS,iBAAiB,UAAU,EAAE;AAAA,EACxC,EAAE,MAAM,MAAM,MAAS;AAEvB,SAAO;AAAA,IACL,MAAM;AAAA,IACN,YAAY,OAAO,eAAe,WAAW,aAAa;AAAA,IAC1D;AAAA,EACF;AACF;AAMA,eAAe,mBACb,MACuE;AACvE,QAAM,UAAwE,CAAC;AAE/E,MAAI;AACF,UAAM,WAAW,MAAM,KAAK,QAAQ,aAAa,EAAE,IAAI;AACvD,eAAW,MAAM,SAAS,MAAM,GAAG,CAAC,GAAG;AACrC,UAAI;AACF,cAAM,SAAS,MAAM,GAAG,UAAU,EAAE,MAAM,MAAM,EAAE;AAClD,cAAM,GAAG,MAAM;AACf,cAAM,cAAc,MAAM,GAAG;AAC7B,cAAM,qBAAqB,MAAM,GAAG,aAAa,WAAW,EAAE,MAAM,MAAM,IAAI;AAC9E,gBAAQ,KAAK,EAAE,QAAQ,OAAO,KAAK,GAAG,mBAAmB,CAAC;AAG1D,cAAM,GAAG,MAAM,EAAE,MAAM,MAAM,MAAS;AACtC,cAAM,cAAc,MAAM,GAAG;AAAA,MAC/B,QAAQ;AAAA,MAER;AAAA,IACF;AAAA,EACF,QAAQ;AAAA,EAER;AAEA,SAAO;AACT;AAMA,eAAsB,0BAA0B,MAA6C;AAE3F,MAAI,UAAuB,CAAC;AAC5B,MAAI;AACF,cAAU,MAAM,gBAAgB,IAAI;AAAA,EACtC,QAAQ;AAAA,EAER;AAGA,MAAI,eAA8B,CAAC;AACnC,MAAI;AACF,mBAAe,MAAM,kBAAkB,MAAM,OAAO;AAAA,EACtD,QAAQ;AAAA,EAER;AAGA,MAAI,aAAgC,EAAE,MAAM,QAAQ,OAAO,CAAC,EAAE;AAC9D,MAAI;AACF,UAAM,OAAO,MAAM,qBAAqB,IAAI;AAE5C,QAAI,SAAS,YAAY;AACvB,mBAAa,MAAM,wBAAwB,IAAI;AAAA,IACjD,WAAW,SAAS,aAAa;AAE/B,mBAAa,MAAM,sBAAsB,IAAI;AAC7C,mBAAa,EAAE,GAAG,YAAY,MAAM,YAAY;AAAA,IAClD,WAAW,SAAS,mBAAmB;AACrC,mBAAa,MAAM,sBAAsB,IAAI;AAAA,IAC/C,OAAO;AAEL,YAAM,YAAY,MAAM,gBAAgB,IAAI;AAC5C,mBAAa,EAAE,MAAM,QAAQ,OAAO,UAAU,SAAS,IAAI,CAAC,EAAE,SAAS,GAAG,UAAU,CAAC,IAAI,CAAC,EAAE;AAAA,IAC9F;AAAA,EACF,QAAQ;AAAA,EAER;AAGA,MAAI,cAA4E,CAAC;AACjF,MAAI;AACF,kBAAc,MAAM,mBAAmB,IAAI;AAAA,EAC7C,QAAQ;AAAA,EAER;AAEA,SAAO;AAAA,IACL;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,EACF;AACF;",
+  "names": []
+}