@intuned/browser-dev 0.1.15-dev.0 → 0.1.16-dev.0

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"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@intuned/browser-dev",
-  "version": "0.1.15-dev.0",
+  "version": "0.1.16-dev.0",
   "description": "runner package for intuned functions",
   "types": "./dist/index.d.ts",
   "typesVersions": {