us.codecraft.webmagic.pipeline.Pipeline Java Examples

@Ignore("long time")
@Test
public void testStartAndStop() throws InterruptedException {
    Spider spider = Spider.create(new SimplePageProcessor( "http://www.oschina.net/*")).addPipeline(new Pipeline() {
        @Override
        public void process(ResultItems resultItems, Task task) {
            System.out.println(1);
        }
    }).thread(1).addUrl("http://www.oschina.net/");
    spider.start();
    Thread.sleep(10000);
    spider.stop();
    Thread.sleep(10000);
    spider.start();
    Thread.sleep(10000);
}
 

private static void startSpider(Params params) {
    ScriptProcessor pageProcessor = ScriptProcessorBuilder.custom()
            .language(params.getLanguage()).scriptFromFile(params.getScriptFileName()).thread(params.getThread()).build();
    pageProcessor.getSite().setSleepTime(params.getSleepTime());
    pageProcessor.getSite().setRetryTimes(3);
    pageProcessor.getSite().setAcceptStatCode(WMCollections.<Integer>newHashSet(200, 404,403, 500,502));
    Spider spider = Spider.create(pageProcessor).thread(params.getThread());
    spider.clearPipeline().addPipeline(new Pipeline() {
        @Override
        public void process(ResultItems resultItems, Task task) {

        }
    });
    if (params.getUrls() == null || params.getUrls().size() == 0) {
        System.err.println("Need at least one argument");
        System.out.println("Usage: java -jar webmagic.jar [-l language] -f script file [-t threadnum] [-s sleep time] url1 [url2 url3]");
        System.exit(-1);
    }
    for (String url : params.getUrls()) {
        spider.addUrl(url);
    }
    spider.run();
}
 

public void close() {
    destroyEach(downloader);
    destroyEach(pageProcessor);
    destroyEach(scheduler);
    for (Pipeline pipeline : pipelines) {
        destroyEach(pipeline);
    }
    threadPool.shutdown();
}
 

private void onDownloadSuccess(Request request, Page page) {
    if (site.getAcceptStatCode().contains(page.getStatusCode())){
        pageProcessor.process(page);
        extractAndAddRequests(page, spawnUrl);
        if (!page.getResultItems().isSkip()) {
            for (Pipeline pipeline : pipelines) {
                pipeline.process(page.getResultItems(), this);
            }
        }
    } else {
        logger.info("page status code error, page {} , code: {}", request.getUrl(), page.getStatusCode());
    }
    sleep(site.getSleepTime());
    return;
}
 

@Test
public void test_github() throws Exception {
    Spider.create(new GithubRepoPageProcessor()).addPipeline(new Pipeline() {
        @Override
        public void process(ResultItems resultItems, Task task) {
            assertThat(((String) resultItems.get("name")).trim()).isEqualTo("webmagic");
            assertThat(((String) resultItems.get("author")).trim()).isEqualTo("code4craft");
        }
    }).setDownloader(new MockGithubDownloader()).test("https://github.com/code4craft/webmagic");
}
 

@Test
public void test() {
    OOSpider.create(new GithubRepoProcessor()).addPipeline(new Pipeline() {
        @Override
        public void process(ResultItems resultItems, Task task) {
            Assert.assertEquals("78",((String)resultItems.get("star")).trim());
            Assert.assertEquals("65",((String)resultItems.get("fork")).trim());
        }
    }).setDownloader(new MockGithubDownloader()).test("https://github.com/code4craft/webmagic");
}
 

public List<Pipeline> getPipelineList() {
    return pipelineList;
}
 

public CommonSpider setPipelineList(List<Pipeline> pipelineList) {
    this.pipelineList = pipelineList;
    return this;
}
 

public List<Pipeline> getPipelineList() {
    return pipelineList;
}
 

public CommonSpider setPipelineList(List<Pipeline> pipelineList) {
    this.pipelineList = pipelineList;
    return this;
}
 

/**
 * add a pipeline for Spider
 *
 * @param pipeline pipeline
 * @return this
 * @see Pipeline
 * @since 0.2.1
 */
public Spider addPipeline(Pipeline pipeline) {
    checkIfRunning();
    this.pipelines.add(pipeline);
    return this;
}
 

/**
 * set pipelines for Spider
 *
 * @param pipelines pipelines
 * @return this
 * @see Pipeline
 * @since 0.4.1
 */
public Spider setPipelines(List<Pipeline> pipelines) {
    checkIfRunning();
    this.pipelines = pipelines;
    return this;
}
 

/**
 * clear the pipelines set
 *
 * @return this
 */
public Spider clearPipeline() {
    pipelines = new ArrayList<Pipeline>();
    return this;
}