Updated URLCharityScraper

Changed driver and wait initialization to be for each method instead of constructor. Added a method for getting categories from charity URL.
cathrkri · Mar 24, 2026 · bc0cc72 · bc0cc72
1 parent 30a5f0e
commit bc0cc72
Showing 1 changed file with 34 additions and 10 deletions.
diff --git a/...mehelpapplication/src/main/java/ntnu/systemutvikling/team6/scraper/URLCharityScraper.java b/...mehelpapplication/src/main/java/ntnu/systemutvikling/team6/scraper/URLCharityScraper.java
@@ -1,6 +1,7 @@
 package ntnu.systemutvikling.team6.scraper;
 
 import java.time.Duration;
+import java.util.ArrayList;
 import java.util.List;
 import org.openqa.selenium.By;
 import org.openqa.selenium.WebDriver;
@@ -12,8 +13,6 @@
 
 public class URLCharityScraper {
   ChromeOptions options;
-  WebDriver driver;
-  WebDriverWait wait;
 
   public URLCharityScraper() {
     this.options = new ChromeOptions();
@@ -22,20 +21,18 @@ public URLCharityScraper() {
     options.addArguments("--disable-gpu");
     options.addArguments("--no-sandbox");
     options.addArguments("--disable-dev-shm-usage");
-
-    this.driver = new ChromeDriver(options);
-
-    this.wait = new WebDriverWait(driver, Duration.ofSeconds(30));
   }
 
   public boolean updateDescription() {
+      WebDriver driver = new ChromeDriver(options);
+      WebDriverWait wait = new WebDriverWait(driver, Duration.ofSeconds(30));
     StringBuilder descriptionString = new StringBuilder();
 
     try {
       // URL for scraping approved organizations
       driver.get("https://www.innsamlingskontrollen.no/organisasjoner/anna-ministries/");
 
-      this.wait.until(
+      wait.until(
           ExpectedConditions.numberOfElementsToBeMoreThan(By.cssSelector(".information div"), 0));
 
       List<WebElement> description = driver.findElements(By.cssSelector(".information div"));
@@ -50,7 +47,7 @@ public boolean updateDescription() {
       if (!doesReadMoreExist.isEmpty()) {
         WebElement descReadMore = driver.findElement(By.cssSelector("a.read-more"));
         descReadMore.click();
-        this.wait.until(ExpectedConditions.visibilityOfElementLocated(By.cssSelector(".extra-info")));
+        wait.until(ExpectedConditions.visibilityOfElementLocated(By.cssSelector(".extra-info")));
 
         List<WebElement> extraDescription = driver.findElements(By.cssSelector(".extra-info"));
 
@@ -66,19 +63,46 @@ public boolean updateDescription() {
   }
 
   public boolean updateLogo() {
+      WebDriver driver = new ChromeDriver(options);
+      WebDriverWait wait = new WebDriverWait(driver, Duration.ofSeconds(30));
       String logoURL = "";
 
       try {
           driver.get("https://www.innsamlingskontrollen.no/organisasjoner/anna-ministries/");
 
-          this.wait.until(ExpectedConditions.visibilityOfElementLocated(By.cssSelector(".logo > img")));
+          wait.until(ExpectedConditions.visibilityOfElementLocated(By.cssSelector(".logo > img")));
 
           WebElement logo = driver.findElement(By.cssSelector(".logo > img"));
 
           logoURL =  logo.getAttribute("src");
       } finally {
-          driver.close();
+          driver.quit();
       }
     return true;
   }
+
+  public boolean updateCategories() {
+      WebDriver driver = new ChromeDriver(options);
+      WebDriverWait wait = new WebDriverWait(driver, Duration.ofSeconds(30));
+      List<String> categoriesList = new ArrayList<>();
+
+      try {
+          driver.get("https://www.innsamlingskontrollen.no/organisasjoner/anna-ministries/");
+
+          wait.until(ExpectedConditions.visibilityOfElementLocated(By.cssSelector(".tag-label")));
+
+          List<WebElement> categories = driver.findElements(By.cssSelector(".tag-label"));
+
+          for (WebElement element : categories) {
+              categoriesList.add(element.getText());
+          }
+
+      } finally {
+          driver.quit();
+      }
+
+      System.out.println(categoriesList);
+
+      return true;
+  }
 }