java/project/BookStrategy.java

package com.yyt.moviecrawler.strategy;

import com.yyt.moviecrawler.model.Book;
import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import org.jsoup.select.Elements;

import java.io.IOException;
import java.util.ArrayList;
import java.util.List;

public class BookStrategy {
    public List<Book> crawl(int limit) {
        List<Book> books = new ArrayList<>();
        try {
            Document doc = Jsoup.connect("http://books.toscrape.com/").get();
            Elements bookElements = doc.select("article.product_pod");

            for (int i = 0; i < Math.min(limit, bookElements.size()); i++) {
                Element el = bookElements.get(i);
                String title = el.select("h3 a").attr("title");
                double price = Double.parseDouble(el.select(".price_color").text().replace("£", ""));
                int star = getStarRating(el.select(".star-rating").attr("class"));
                String category = "Books to Scrape";

                books.add(new Book(title, price, star, category));
            }
        } catch (IOException e) {
            e.printStackTrace();
        }
        return books;
    }

    private int getStarRating(String className) {
        if (className.contains("One")) return 1;
        if (className.contains("Two")) return 2;
        if (className.contains("Three")) return 3;
        if (className.contains("Four")) return 4;
        if (className.contains("Five")) return 5;
        return 0;
    }
}