@intuned/browser-dev 0.1.15-dev.1 → 0.1.16-dev.1

package/dist/optimized-extractors/listExtractionHelpers/runAiExtraction.js

@@ -21,6 +21,24 @@ var _findTableHeaders = require("../common/findTableHeaders");
 var _Logger = require("../../common/Logger");
 var _utils = require("../common/matching/utils");
 function _interopRequireWildcard(e, t) { if ("function" == typeof WeakMap) var r = new WeakMap(), n = new WeakMap(); return (_interopRequireWildcard = function (e, t) { if (!t && e && e.__esModule) return e; var o, i, f = { __proto__: null, default: e }; if (null === e || "object" != typeof e && "function" != typeof e) return f; if (o = t ? n : r) { if (o.has(e)) return o.get(e); o.set(e, f); } for (const t in e) "default" !== t && {}.hasOwnProperty.call(e, t) && ((i = (o = Object.defineProperty) && Object.getOwnPropertyDescriptor(e, t)) && (i.get || i.set) ? o(f, t, i) : f[t] = e[t]); return f; })(e, t); }
+function getCommonAncestorXpath(xpaths) {
+  if (xpaths.length === 0) return null;
+  const segmented = xpaths.map(xpath => xpath.split("/"));
+  const minLen = Math.min(...segmented.map(segments => segments.length));
+  const common = [];
+  for (let i = 0; i < minLen; i++) {
+    const segment = segmented[0][i];
+    if (segmented.every(segments => segments[i] === segment)) {
+      common.push(segment);
+    } else {
+      break;
+    }
+  }
+  if (common.length > 1 && xpaths.some(xpath => xpath === common.join("/"))) {
+    common.pop();
+  }
+  return common.length > 0 ? common.join("/") : null;
+}
 async function runAiExtraction(params) {
   return handleNewAiExtraction(params);
 }
@@ -82,7 +100,8 @@ async function handleNewAiExtraction(params) {
           acc[key] = {
             matchText: bestMatch.matchText,
             matchXpath: bestMatch.matchXpath,
-            matchType: bestMatch.matchType
+            matchType: bestMatch.matchType,
+            sourceText: bestMatch.sourceText
           };
         } else {
           _Logger.logger.debug(`value "${value}" for key "${key}" in row ${i + 1} does not have any matches in the page html, dropped for hallucination protection`);
@@ -100,26 +119,15 @@ async function handleNewAiExtraction(params) {
     })));
     let containerPath = null;
     let fullContainerXpath = null;
-    if (resultValues.length > 0 && resultValues[0].result[primaryPropertyName]) {
-      const primaryXpath = resultValues[0].result[primaryPropertyName].matchXpath;
-      if (primaryXpath) {
-        const parentXpath = await pageAndSearchRegion.page.evaluate(xpath => {
-          var _window$__INTUNED__;
-          const result = document.evaluate(xpath, document, null, XPathResult.FIRST_ORDERED_NODE_TYPE, null);
-          const element = result.singleNodeValue;
-          if (!element || !element.parentElement) return null;
-          if ((_window$__INTUNED__ = window.__INTUNED__) !== null && _window$__INTUNED__ !== void 0 && _window$__INTUNED__.getElementXPath) {
-            return window.__INTUNED__.getElementXPath(element.parentElement);
-          }
-          return null;
-        }, primaryXpath);
-        if (parentXpath) {
-          fullContainerXpath = parentXpath;
-          if (hasSearchRegionContainer) {
-            containerPath = await (0, _getRelativeContainerXpathSelector.getRelativeContainerXpathSelector)(pageAndSearchRegion.searchRegion, parentXpath);
-          } else {
-            containerPath = parentXpath;
-          }
+    if (resultValues.length > 0) {
+      const propertyXpaths = Object.values(resultValues[0].result).map(value => value.matchXpath).filter(xpath => !!xpath);
+      const itemContainerXpath = getCommonAncestorXpath(propertyXpaths);
+      if (itemContainerXpath) {
+        fullContainerXpath = itemContainerXpath;
+        if (hasSearchRegionContainer) {
+          containerPath = await (0, _getRelativeContainerXpathSelector.getRelativeContainerXpathSelector)(pageAndSearchRegion.searchRegion, itemContainerXpath);
+        } else {
+          containerPath = itemContainerXpath;
         }
       }
     }
@@ -201,7 +209,7 @@ async function splitDomAndExtractData({
     tableLocater
   } = await (0, _tablesUtils.isListTable)(listItemsContainerLocator, itemsSimplifiedHtml);
   const tableAsJsonArray = isTable ? await (0, _tablesUtils.createJsonFromTable)(pageAndSearchRegion.page) : [];
-  const tableHeaders = tableLocater ? await (0, _findTableHeaders.getTableHeadersUsingAi)(tableLocater, identifier) : undefined;
+  const tableHeaders = tableLocater ? await (0, _findTableHeaders.getTableHeadersUsingAi)(tableLocater, strategy.model) : undefined;
   if (tableHeaders && tableHeaders.isErr()) {
     return (0, _neverthrow.err)(tableHeaders.error);
   }
@@ -240,7 +248,8 @@ async function splitDomAndExtractData({
         acc[key] = {
           matchText: bestMatch.matchText,
           matchXpath: bestMatch.matchXpath,
-          matchType: bestMatch.matchType
+          matchType: bestMatch.matchType,
+          sourceText: bestMatch.sourceText
         };
       } else {
         _Logger.logger.debug(`value "${value}" for key "${key}" in row ${i + 1} does not have any matches in the item's html, dropped for hallucination protection`);

package/dist/optimized-extractors/listExtractionHelpers/utils/extractPropertiesUsingGPTFromArray.js

@@ -44,7 +44,7 @@ async function extractPropertiesUsingGPT({
   }
   if (possibleTableHeaderOrFooter) {
     const content = text ?? image;
-    const isHeader = await (0, _isTableHeaderOrFooter.isTableHeaderOrFooter)(content);
+    const isHeader = await (0, _isTableHeaderOrFooter.isTableHeaderOrFooter)(content, strategy.model);
     if (isHeader.isErr()) {
       return (0, _neverthrow.err)(isHeader.error);
     }
@@ -146,7 +146,7 @@ async function extractPropertiesWithHTMLStrategy({
   apiKey
 }) {
   const shouldUseTableData = !!tableAsJsonArray && tableAsJsonArray.length === items.length;
-  const isWeakModel = strategy.model === "claude-3-haiku" || strategy.model == "gpt3.5-turbo";
+  const isWeakModel = strategy.model.includes("haiku") || strategy.model.includes("turbo");
   const averageItemLength = items.reduce((sum, item) => {
     if (item.type !== "text") return sum;
     return sum + (0, _extractionHelpers.compressStringSpaces)(item.text).length;

package/dist/optimized-extractors/objectExtractionHelpers/AIExtractors.js

@@ -59,7 +59,8 @@ async function runAIExtraction(pageAndSearchRegion, extractionInfo, snapshot, st
       xpathMapping[propertyName] = {
         matchXpath: bestMatch.matchXpath,
         matchText: bestMatch.matchText,
-        matchType: bestMatch.matchType
+        matchType: bestMatch.matchType,
+        sourceText: bestMatch.sourceText
       };
     } else {
       _Logger.logger.debug(`Property ${propertyName} not found in the page HTML, dropped for hallucination protection`);

package/dist/optimized-extractors/objectExtractionHelpers/__tests__/testObjectExtractorFromLocator.spec.js

@@ -2,7 +2,9 @@
 
 var _extendedTest = require("../../../common/extendedTest");
 var _ = require("../..");
+var _neverthrow = require("neverthrow");
 var _uuid = require("uuid");
+function _interopRequireWildcard(e, t) { if ("function" == typeof WeakMap) var r = new WeakMap(), n = new WeakMap(); return (_interopRequireWildcard = function (e, t) { if (!t && e && e.__esModule) return e; var o, i, f = { __proto__: null, default: e }; if (null === e || "object" != typeof e && "function" != typeof e) return f; if (o = t ? n : r) { if (o.has(e)) return o.get(e); o.set(e, f); } for (const t in e) "default" !== t && {}.hasOwnProperty.call(e, t) && ((i = (o = Object.defineProperty) && Object.getOwnPropertyDescriptor(e, t)) && (i.get || i.set) ? o(f, t, i) : f[t] = e[t]); return f; })(e, t); }
 const productTemplate = `
 <div class="product-page">
   <div class="product-info">
@@ -104,4 +106,210 @@ _extendedTest.describe.skip("Object Extractor Caching Tests", () => {
       console.log("All cache behavior tests completed successfully!");
     });
   });
+});
+async function mockExtractStructuredDataUsingAi(result) {
+  const commonModule = await Promise.resolve().then(() => _interopRequireWildcard(require("../../common")));
+  return _extendedTest.vi.spyOn(commonModule, "extractStructuredDataUsingAi").mockResolvedValue((0, _neverthrow.ok)({
+    result
+  }));
+}
+async function useInMemoryCache() {
+  const cacheModule = await Promise.resolve().then(() => _interopRequireWildcard(require("../../../intunedServices/cache")));
+  const store = new Map();
+  _extendedTest.vi.spyOn(cacheModule.cache, "get").mockImplementation(async key => store.has(key) ? store.get(key) : null);
+  _extendedTest.vi.spyOn(cacheModule.cache, "set").mockImplementation(async (key, value) => {
+    store.set(key, value);
+  });
+}
+(0, _extendedTest.describe)("extractObjectFromLocator - cache round-trip (real matcher, mocked AI)", () => {
+  (0, _extendedTest.afterEach)(() => {
+    _extendedTest.vi.restoreAllMocks();
+  });
+  (0, _extendedTest.test)("reuses the cache on a second identical call when the match was partial (PH: prefix)", async ({
+    page
+  }) => {
+    const label = `contact-roundtrip-${(0, _uuid.v4)()}`;
+    const aiResult = {
+      contact_information_name: "Andreas Jansson",
+      contact_information_phone_number: "(727) 471-4768",
+      contact_information_email: "andreas.jansson@flowbird.group"
+    };
+    const aiSpy = await mockExtractStructuredDataUsingAi(aiResult);
+    await useInMemoryCache();
+    await page.setContent(`
+      <div class="vendor">
+        <h2>Contact Information</h2>
+        <p class="c-name">Andreas Jansson</p>
+        <p class="c-phone">PH: (727) 471-4768</p>
+        <p class="c-email">andreas.jansson@flowbird.group</p>
+      </div>
+    `);
+    const options = {
+      entityName: "supplier_contacts",
+      label,
+      entitySchema: {
+        type: "object",
+        required: ["contact_information_email"],
+        properties: {
+          contact_information_name: {
+            type: "string",
+            description: "the contact person name"
+          },
+          contact_information_phone_number: {
+            type: "string",
+            description: "the contact phone number"
+          },
+          contact_information_email: {
+            type: "string",
+            description: "the contact email"
+          }
+        }
+      },
+      strategy: {
+        model: "claude-sonnet-4-20250514",
+        type: "HTML"
+      },
+      variantKey: label
+    };
+    const first = await (0, _.extractObjectFromLocator)(page.locator("div.vendor"), options);
+    (0, _extendedTest.expect)(first).toEqual(aiResult);
+    (0, _extendedTest.expect)(aiSpy).toHaveBeenCalledTimes(1);
+    const second = await (0, _.extractObjectFromLocator)(page.locator("div.vendor"), options);
+    (0, _extendedTest.expect)(second).toEqual(first);
+    (0, _extendedTest.expect)(aiSpy).toHaveBeenCalledTimes(1);
+  });
+  (0, _extendedTest.test)("reuses the cache when fields sit in separate direct text nodes of one element (br-separated block)", async ({
+    page
+  }) => {
+    const label = `vendor-block-${(0, _uuid.v4)()}`;
+    const vendorAiResult = {
+      primary_vendor_name: "V00000110 - International Cleaning Services, Inc.",
+      primary_vendor_contact_email: "monaco.intl@sbcglobal.net",
+      primary_vendor_contact_phone: "(630) 904-4118"
+    };
+    const aiSpy = await mockExtractStructuredDataUsingAi(vendorAiResult);
+    await useInMemoryCache();
+    await page.setContent(`
+      <table><tbody><tr><td class="vendor">
+        <a href="#">V00000110 - International Cleaning Services, Inc.</a>
+        2415 Comstock Court<br/>
+        Naperville, IL 60564<br/>
+        US<br/>
+        Email: monaco.intl@sbcglobal.net<br/>
+        FAX: (630) 904-4118
+      </td></tr></tbody></table>
+    `);
+    const options = {
+      entityName: "supplier_contacts",
+      label,
+      entitySchema: {
+        type: "object",
+        required: ["primary_vendor_name"],
+        properties: {
+          primary_vendor_name: {
+            type: "string",
+            description: "vendor id - company name"
+          },
+          primary_vendor_contact_email: {
+            type: "string",
+            description: "vendor email"
+          },
+          primary_vendor_contact_phone: {
+            type: "string",
+            description: "vendor phone"
+          }
+        }
+      },
+      strategy: {
+        model: "claude-sonnet-4-20250514",
+        type: "HTML"
+      },
+      variantKey: label
+    };
+    const first = await (0, _.extractObjectFromLocator)(page.locator("td.vendor"), options);
+    (0, _extendedTest.expect)(first).toEqual(vendorAiResult);
+    (0, _extendedTest.expect)(aiSpy).toHaveBeenCalledTimes(1);
+    const second = await (0, _.extractObjectFromLocator)(page.locator("td.vendor"), options);
+    (0, _extendedTest.expect)(second).toEqual(first);
+    (0, _extendedTest.expect)(aiSpy).toHaveBeenCalledTimes(1);
+  });
+  (0, _extendedTest.test)("reuses the cache for a link whose href is stored resolved but read raw (relative href)", async ({
+    page
+  }) => {
+    const label = `href-${(0, _uuid.v4)()}`;
+    const hrefAiResult = {
+      attachment_href: "https://example.com/files/annual-report.pdf",
+      attachment_name: "Annual Report"
+    };
+    const aiSpy = await mockExtractStructuredDataUsingAi(hrefAiResult);
+    await useInMemoryCache();
+    await page.setContent(`<!DOCTYPE html><html><head><base href="https://example.com/files/"></head><body>` + `<div class="doc"><a href="annual-report.pdf">Annual Report</a></div>` + `</body></html>`);
+    const options = {
+      entityName: "attachment",
+      label,
+      entitySchema: {
+        type: "object",
+        required: ["attachment_href"],
+        properties: {
+          attachment_href: {
+            type: "string",
+            description: "the file url (href)"
+          },
+          attachment_name: {
+            type: "string",
+            description: "the file name"
+          }
+        }
+      },
+      strategy: {
+        model: "claude-sonnet-4-20250514",
+        type: "HTML"
+      },
+      variantKey: label
+    };
+    const first = await (0, _.extractObjectFromLocator)(page.locator("div.doc"), options);
+    (0, _extendedTest.expect)(first).toEqual(hrefAiResult);
+    (0, _extendedTest.expect)(aiSpy).toHaveBeenCalledTimes(1);
+    const second = await (0, _.extractObjectFromLocator)(page.locator("div.doc"), options);
+    (0, _extendedTest.expect)(second).toEqual(first);
+    (0, _extendedTest.expect)(aiSpy).toHaveBeenCalledTimes(1);
+  });
+  (0, _extendedTest.test)("reuses the cache for a FUZZY match (AI value differs slightly from the DOM)", async ({
+    page
+  }) => {
+    const label = `fuzzy-${(0, _uuid.v4)()}`;
+    const fuzzyAiResult = {
+      company_name: "Internationl Cleaning Servces"
+    };
+    const aiSpy = await mockExtractStructuredDataUsingAi(fuzzyAiResult);
+    await useInMemoryCache();
+    await page.setContent(`<div class="vendor"><span class="name">International Cleaning Services</span></div>`);
+    const options = {
+      entityName: "vendor",
+      label,
+      entitySchema: {
+        type: "object",
+        required: ["company_name"],
+        properties: {
+          company_name: {
+            type: "string",
+            description: "the company name"
+          }
+        }
+      },
+      strategy: {
+        model: "claude-sonnet-4-20250514",
+        type: "HTML"
+      },
+      variantKey: label
+    };
+    const first = await (0, _.extractObjectFromLocator)(page.locator("div.vendor"), options);
+    (0, _extendedTest.expect)(first).toEqual({
+      company_name: "International Cleaning Services"
+    });
+    (0, _extendedTest.expect)(aiSpy).toHaveBeenCalledTimes(1);
+    const second = await (0, _.extractObjectFromLocator)(page.locator("div.vendor"), options);
+    (0, _extendedTest.expect)(second).toEqual(first);
+    (0, _extendedTest.expect)(aiSpy).toHaveBeenCalledTimes(1);
+  });
 });

package/dist/optimized-extractors/objectExtractionHelpers/dynamicObjectExtractor.js

@@ -88,11 +88,13 @@ async function dynamicObjectExtractor(page, identifier, options) {
   Object.entries(xpathMappingFromAI).forEach(([_propertyName, {
     matchXpath,
     matchText,
-    matchType
+    matchType,
+    sourceText
   }]) => {
     xpathMapping[matchText] = [{
       xpath: matchXpath,
-      matchType
+      matchType,
+      sourceText
     }];
   });
   const resultsToCache = {

package/dist/optimized-extractors/validators.js

@@ -5,22 +5,21 @@ Object.defineProperty(exports, "__esModule", {
 });
 exports.strategySchema = exports.simpleObjectJsonSchema = exports.simpleArrayItemJsonSchema = exports.extractObjectOptimizedInputSchema = exports.extractArrayOptimizedInputSchema = void 0;
 var _zod = require("zod");
-var _aiModelsValidation = require("./types/aiModelsValidation");
 const htmlStrategySchema = _zod.z.object({
-  model: _zod.z.enum(_aiModelsValidation.SUPPORTED_TEXT_MODELS, {
+  model: _zod.z.string({
     required_error: "strategy model is required",
     invalid_type_error: "strategy model is invalid"
-  }),
+  }).min(1, "strategy model is required"),
   type: _zod.z.literal("HTML", {
     required_error: "strategy type is required",
     invalid_type_error: "strategy type is invalid"
   })
 });
 const imageStrategySchema = _zod.z.object({
-  model: _zod.z.enum(_aiModelsValidation.SUPPORTED_VISION_MODELS, {
+  model: _zod.z.string({
     required_error: "strategy model is required",
     invalid_type_error: "strategy model is invalid"
-  }),
+  }).min(1, "strategy model is required"),
   type: _zod.z.literal("IMAGE", {
     required_error: "strategy type is required",
     invalid_type_error: "strategy type is invalid"
@@ -114,7 +113,7 @@ const strategySchema = exports.strategySchema = _zod.z.union([htmlStrategySchema
     };
   }
 }).optional().default({
-  model: "claude-3-haiku",
+  model: "claude-haiku-4-5-20251001",
   type: "HTML"
 });
 const labelSchema = _zod.z.string({

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@intuned/browser-dev",
-  "version": "0.1.15-dev.1",
+  "version": "0.1.16-dev.1",
   "description": "runner package for intuned functions",
   "types": "./dist/index.d.ts",
   "typesVersions": {

package/dist/optimized-extractors/types/aiModelsValidation.js

@@ -1,45 +0,0 @@
-"use strict";
-
-Object.defineProperty(exports, "__esModule", {
-  value: true
-});
-exports.SUPPORTED_VISION_MODELS = exports.SUPPORTED_TEXT_MODELS = exports.SUPPORTED_GPT_MODELS = exports.SUPPORTED_GOOGLE_MODELS = exports.SUPPORTED_CLAUDE_MODELS = exports.MODELS_MAPPINGS = exports.MAX_TOKENS_OVERRIDES = exports.GPT_MODELS_MAPPINGS = exports.GOOGLE_MODELS_MAPPINGS = exports.CLAUDE_VISION_SUPPORTED_MODELS = exports.CLAUDE_ONLY_TEXT_MODELS = exports.CLAUDE_MODELS_MAPPINGS = void 0;
-const CLAUDE_ONLY_TEXT_MODELS = exports.CLAUDE_ONLY_TEXT_MODELS = ["claude-3-5-haiku", "claude-3-5-haiku-20241022"];
-const CLAUDE_VISION_SUPPORTED_MODELS = exports.CLAUDE_VISION_SUPPORTED_MODELS = ["claude-3-haiku", "claude-3-haiku-20240307", "claude-3.5-sonnet", "claude-3-5-sonnet-20240620", "claude-3-5-sonnet-20241022", "claude-opus-4", "claude-opus-4-20250514", "claude-sonnet-4", "claude-sonnet-4-20250514"];
-const SUPPORTED_CLAUDE_MODELS = exports.SUPPORTED_CLAUDE_MODELS = [...CLAUDE_ONLY_TEXT_MODELS, ...CLAUDE_VISION_SUPPORTED_MODELS];
-const CLAUDE_MODELS_MAPPINGS = exports.CLAUDE_MODELS_MAPPINGS = {
-  "claude-3-haiku": "claude-3-haiku-20240307",
-  "claude-3-5-haiku": "claude-3-5-haiku-20241022",
-  "claude-3.5-sonnet": "claude-3-5-sonnet-20241022",
-  "claude-opus-4": "claude-opus-4-20250514",
-  "claude-sonnet-4": "claude-sonnet-4-20250514"
-};
-const GPT_ONLY_TEXT_GPT_MODELS = ["gpt3.5-turbo", "gpt-3.5-turbo-0125"];
-const GPT_VISION_SUPPORTED_MODELS = ["gpt4-turbo", "gpt-4-turbo-2024-04-09", "gpt-4o", "gpt-4o-2024-05-13", "gpt-4o-mini", "gpt-4o-mini-2024-07-18"];
-const SUPPORTED_GPT_MODELS = exports.SUPPORTED_GPT_MODELS = [...GPT_ONLY_TEXT_GPT_MODELS, ...GPT_VISION_SUPPORTED_MODELS];
-const GPT_MODELS_MAPPINGS = exports.GPT_MODELS_MAPPINGS = {
-  "gpt4-turbo": "gpt-4-turbo-2024-04-09",
-  "gpt3.5-turbo": "gpt-3.5-turbo-0125",
-  "gpt-4o": "gpt-4o-2024-05-13",
-  "gpt-4o-mini": "gpt-4o-mini-2024-07-18"
-};
-const SUPPORTED_GOOGLE_MODELS = exports.SUPPORTED_GOOGLE_MODELS = ["gemini-1.5-pro", "gemini-1.5-pro-002", "gemini-1.5-flash-8b", "gemini-1.5-flash-8b-002", "gemini-1.5-flash", "gemini-1.5-flash-002", "gemini-2.0-flash-exp"];
-const GOOGLE_MODELS_MAPPINGS = exports.GOOGLE_MODELS_MAPPINGS = {
-  "gemini-1.5-pro": "gemini-1.5-pro-002",
-  "gemini-1.5-flash-8b": "gemini-1.5-flash-8b-002",
-  "gemini-1.5-flash": "gemini-1.5-flash-002"
-};
-const SUPPORTED_TEXT_MODELS = exports.SUPPORTED_TEXT_MODELS = [...SUPPORTED_CLAUDE_MODELS, ...SUPPORTED_GPT_MODELS, ...SUPPORTED_GOOGLE_MODELS];
-const SUPPORTED_VISION_MODELS = exports.SUPPORTED_VISION_MODELS = [...CLAUDE_VISION_SUPPORTED_MODELS, ...GPT_VISION_SUPPORTED_MODELS, ...SUPPORTED_GOOGLE_MODELS];
-const MODELS_MAPPINGS = exports.MODELS_MAPPINGS = {
-  ...GPT_MODELS_MAPPINGS,
-  ...CLAUDE_MODELS_MAPPINGS,
-  ...GOOGLE_MODELS_MAPPINGS
-};
-const MAX_TOKENS_OVERRIDES = exports.MAX_TOKENS_OVERRIDES = {
-  "claude-3-5-sonnet-20240620": 8192,
-  "gemini-1.5-pro-002": 8192,
-  "gemini-1.5-flash-8b-002": 8192,
-  "gemini-1.5-flash-002": 8192,
-  "gemini-2.0-flash-exp": 8192
-};