fix #596: prevent events search category from over-extracting fields (#606)

- Fix splitLines logic to split by newlines before normalizing whitespace in linkedinSearch.ts and linkedinEvents.ts
- Add e2e assertions to ensure title and date_time fields are reasonably sized

Author: sigvardt

packages/core/src/__tests__/e2e/events.e2e.test.ts

@@ -29,7 +29,9 @@ describe("Events E2E", () => {
     const first = result.results[0]!;
     expect(first.event_id.length).toBeGreaterThan(0);
     expect(first.title.length).toBeGreaterThan(0);
+    expect(first.title.length).toBeLessThan(200);
     expect(typeof first.date_time).toBe("string");
+    expect(first.date_time.length).toBeLessThan(100);
     expect(typeof first.location).toBe("string");
     expect(typeof first.organizer).toBe("string");
     expect(typeof first.attendee_count).toBe("string");

packages/core/src/__tests__/e2e/search.e2e.test.ts

@@ -129,6 +129,8 @@ describe("Search E2E", () => {
     expect(first).toBeDefined();
     if (first) {
       expect(first.title.length).toBeGreaterThan(0);
+      expect(first.title.length).toBeLessThan(200);
+      expect(first.date.length).toBeLessThan(100);
     }
   });
 

packages/core/src/linkedinEvents.ts

@@ -344,7 +344,7 @@ async function extractEventSearchResults(
     return uniqueLinks.slice(0, maxEvents).map((link) => {
       const card = link.closest("li") ?? link.closest("div[data-view-tracking-scope]") ?? link.closest("div.search-result__wrapper") ?? link.parentElement;
 
-      const rawText = normalize((card as HTMLElement)?.innerText ?? link.innerText ?? "");
+      const rawText = (card as HTMLElement)?.innerText ?? link.innerText ?? "";
       const lines = rawText
         .split("\n")
         .map((line) => normalize(line))

packages/core/src/linkedinSearch.ts

@@ -2295,7 +2295,7 @@ export class LinkedInSearchService {
 
             return Array.from(globalThis.document.querySelectorAll("main li, div.search-results-container ul > li, ul.reusable-search__entity-result-list > li"))
               .map((card) => {
-                const lines = normalize((card as HTMLElement).innerText)
+                const lines = ((card as HTMLElement).innerText || "")
                   .split("\n")
                   .map((line) => normalize(line))
                   .filter(Boolean);