java/w10/SpringerLinkStrategy.java

package strategy;

import model.Paper;
import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import org.jsoup.select.Elements;
import utils.Utils;
import java.util.ArrayList;
import java.util.List;

public class SpringerLinkStrategy extends AbstractCrawlerStrategy {
    @Override
    public String getPlatformName() {
        return "Springer Link";
    }
    
    @Override
    public boolean supportsUrl(String url) {
        return url != null && url.contains("link.springer.com");
    }
    
    @Override
    protected List<Paper> fetchPapers(String url, int count) throws Exception {
        List<Paper> papers = new ArrayList<>();
        System.out.println("=== 开始使用Springer Link获取论文 ===");
        
        addDelay(2000, 3000);
        
        String html = Utils.sendGetRequest(url);
        if (html.isEmpty()) return papers;
        
        Document doc = Jsoup.parse(html);
        
        Elements paperElements = doc.select(".result-item");
        
        int collected = 0;
        for (Element element : paperElements) {
            if (collected >= count) break;
            
            try {
                Element titleElement = element.selectFirst("h2 a");
                String title = titleElement != null ? titleElement.text() : "";
                
                String paperUrl = titleElement != null ? titleElement.attr("href") : "";
                if (!paperUrl.startsWith("http")) {
                    paperUrl = "https://link.springer.com" + paperUrl;
                }
                
                Element authorsElement = element.selectFirst(".authors");
                String authors = authorsElement != null ? authorsElement.text() : "";
                
                if (title.length() < 5 || paperUrl.isEmpty()) continue;
                
                papers.add(new Paper(title, authors, "", paperUrl, getPlatformName()));
                collected++;
            } catch (Exception e) {
                continue;
            }
        }
        
        return papers;
    }
}