java/project/crawler/CrawlerFactory.java

package com.crawler.crawler;

import com.crawler.crawler.impl.*;
import com.crawler.exception.InvalidUrlException;
import java.util.LinkedHashMap;
import java.util.Map;
import java.util.regex.Pattern;

public class CrawlerFactory {
    private static CrawlerFactory instance;
    private Map<String, Pattern> crawlerPatterns;

    private CrawlerFactory() {
        crawlerPatterns = new LinkedHashMap<>();
        initPatterns();
    }

    public static CrawlerFactory getInstance() {
        if (instance == null) {
            instance = new CrawlerFactory();
        }
        return instance;
    }

    private void initPatterns() {
        crawlerPatterns.put("MountBladeCrawler", 
            Pattern.compile(".*mountblade\\.com\\.cn.*", Pattern.CASE_INSENSITIVE));
        crawlerPatterns.put("HunanUniversityNewsCrawler", 
            Pattern.compile(".*news\\.hnu\\.edu\\.cn.*", Pattern.CASE_INSENSITIVE));
        crawlerPatterns.put("HunanUniversityCrawler", 
            Pattern.compile(".*hnu\\.edu\\.cn.*", Pattern.CASE_INSENSITIVE));
        crawlerPatterns.put("ChinaWeatherCrawler", 
            Pattern.compile(".*weather\\.com\\.cn.*", Pattern.CASE_INSENSITIVE));
        crawlerPatterns.put("ExampleCrawler", 
            Pattern.compile(".*", Pattern.CASE_INSENSITIVE));
    }

    public Crawler createCrawler(String url) {
        validateUrl(url);

        for (Map.Entry<String, Pattern> entry : crawlerPatterns.entrySet()) {
            if (entry.getValue().matcher(url).matches()) {
                return createCrawlerByName(entry.getKey());
            }
        }

        return new ExampleCrawler();
    }

    private void validateUrl(String url) {
        if (url == null || url.isEmpty()) {
            throw new InvalidUrlException("URL不能为空", url);
        }
        
        if (!url.startsWith("http://") && !url.startsWith("https://")) {
            throw new InvalidUrlException("URL格式无效，必须以http://或https://开头", url);
        }
    }

    private Crawler createCrawlerByName(String crawlerName) {
        switch (crawlerName) {
            case "MountBladeCrawler":
                return new MountBladeCrawler();
            case "HunanUniversityNewsCrawler":
                return new HunanUniversityNewsCrawler();
            case "HunanUniversityCrawler":
                return new HunanUniversityCrawler();
            case "ChinaWeatherCrawler":
                return new ChinaWeatherCrawler();
            case "ExampleCrawler":
            default:
                return new ExampleCrawler();
        }
    }

    public String getCrawlerName(String url) {
        if (url == null || url.isEmpty()) {
            return "ExampleCrawler";
        }

        for (Map.Entry<String, Pattern> entry : crawlerPatterns.entrySet()) {
            if (entry.getValue().matcher(url).matches()) {
                return entry.getKey();
            }
        }

        return "ExampleCrawler";
    }

    public boolean isUrlSupported(String url) {
        if (url == null || url.isEmpty()) {
            return false;
        }
        
        for (Pattern pattern : crawlerPatterns.values()) {
            if (pattern.matcher(url).matches()) {
                return true;
            }
        }
        
        return true;
    }
}