Java Code Examples for us.codecraft.webmagic.Page#setRawText()

The following examples show how to use us.codecraft.webmagic.Page#setRawText() . You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may check out the related API usage on the sidebar.
Example 1
Source File: HttpClientDownloader.java    From blog-hunter with MIT License 6 votes vote down vote up 
protected Page handleResponse(Request request, String charset, HttpResponse httpResponse, Task task) throws IOException {
    byte[] bytes = IOUtils.toByteArray(httpResponse.getEntity().getContent());
    String contentType = httpResponse.getEntity().getContentType() == null ? "" : httpResponse.getEntity().getContentType().getValue();
    Page page = new Page();
    page.setBytes(bytes);
    if (!request.isBinaryContent()){
        if (charset == null) {
            charset = getHtmlCharset(contentType, bytes);
        }
        page.setCharset(charset);
        page.setRawText(new String(bytes, charset));
    }
    page.setUrl(new PlainText(request.getUrl()));
    page.setRequest(request);
    page.setStatusCode(httpResponse.getStatusLine().getStatusCode());
    page.setDownloadSuccess(true);
    if (responseHeader) {
        page.setHeaders(HttpClientUtils.convertHeaders(httpResponse.getAllHeaders()));
    }
    return page;
}
Example 2
Source File: HttpClientDownloader.java    From plumemo with Apache License 2.0 6 votes vote down vote up 
protected Page handleResponse(Request request, String charset, HttpResponse httpResponse, Task task) throws IOException {
    byte[] bytes = IOUtils.toByteArray(httpResponse.getEntity().getContent());
    String contentType = httpResponse.getEntity().getContentType() == null ? "" : httpResponse.getEntity().getContentType().getValue();
    Page page = new Page();
    page.setBytes(bytes);
    if (!request.isBinaryContent()) {
        if (charset == null) {
            charset = getHtmlCharset(contentType, bytes);
        }
        page.setCharset(charset);
        page.setRawText(new String(bytes, charset));
    }
    page.setUrl(new PlainText(request.getUrl()));
    page.setRequest(request);
    page.setStatusCode(httpResponse.getStatusLine().getStatusCode());
    page.setDownloadSuccess(true);
    if (responseHeader) {
        page.setHeaders(HttpClientUtils.convertHeaders(httpResponse.getAllHeaders()));
    }
    return page;
}
Example 3
Source File: CasperjsDownloader.java    From spider with GNU General Public License v3.0 6 votes vote down vote up 
@Override
public Page download(Request request, Task task) {
    String html = null;
    Site site = null;
    if (task != null) {
        site = task.getSite();
    }
    try {
        html = casperjs.gatherHtml(new com.gs.spider.model.commons.Request(request.getUrl(), true));
    } catch (Exception e) {
        if (site.getCycleRetryTimes() > 0) {
            return addToCycleRetry(request, site);
        }
        request.putExtra("EXCEPTION", e);
        onError(request);
        return null;
    }
    Page page = new Page();
    page.setRawText(html);
    page.setUrl(new PlainText(request.getUrl()));
    page.setRequest(request);
    onSuccess(request);
    return page;
}
Example 4
Source File: HttpClientDownloader.java    From webmagic with Apache License 2.0 6 votes vote down vote up 
protected Page handleResponse(Request request, String charset, HttpResponse httpResponse, Task task) throws IOException {
    byte[] bytes = IOUtils.toByteArray(httpResponse.getEntity().getContent());
    String contentType = httpResponse.getEntity().getContentType() == null ? "" : httpResponse.getEntity().getContentType().getValue();
    Page page = new Page();
    page.setBytes(bytes);
    if (!request.isBinaryContent()){
        if (charset == null) {
            charset = getHtmlCharset(contentType, bytes);
        }
        page.setCharset(charset);
        page.setRawText(new String(bytes, charset));
    }
    page.setUrl(new PlainText(request.getUrl()));
    page.setRequest(request);
    page.setStatusCode(httpResponse.getStatusLine().getStatusCode());
    page.setDownloadSuccess(true);
    if (responseHeader) {
        page.setHeaders(HttpClientUtils.convertHeaders(httpResponse.getAllHeaders()));
    }
    return page;
}
Example 5
Source File: MockGithubDownloader.java    From webmagic with Apache License 2.0 5 votes vote down vote up 
@Override
public Page download(Request request, Task task) {
    Page page = new Page();
    InputStream resourceAsStream = this.getClass().getResourceAsStream("/html/mock-github.html");
    try {
        page.setRawText(IOUtils.toString(resourceAsStream));
    } catch (IOException e) {
        e.printStackTrace();
    }
    page.setRequest(new Request("https://github.com/code4craft/webmagic"));
    page.setUrl(new PlainText("https://github.com/code4craft/webmagic"));
    return page;
}
Example 6
Source File: PageMocker.java    From webmagic with Apache License 2.0 5 votes vote down vote up 
public Page getMockJsonPage() throws IOException {
    Page page = new Page();
    page.setRawText(IOUtils.toString(PageMocker.class.getClassLoader().getResourceAsStream("json/mock-githubrepo.json")));
    page.setRequest(new Request("https://api.github.com/repos/code4craft/webmagic"));
    page.setUrl(new PlainText("https://api.github.com/repos/code4craft/webmagic"));
    return page;
}
Example 7
Source File: PageMocker.java    From webmagic with Apache License 2.0 5 votes vote down vote up 
public Page getMockPage() throws IOException {
    Page page = new Page();
    page.setRawText(IOUtils.toString(PageMocker.class.getClassLoader().getResourceAsStream("html/mock-webmagic.html")));
    page.setRequest(new Request("http://webmagic.io/list/0"));
    page.setUrl(new PlainText("http://webmagic.io/list/0"));
    return page;
}
Example 8
Source File: ModelPageProcessorTest.java    From webmagic with Apache License 2.0 5 votes vote down vote up 
@Test
public void testMultiModel_should_not_skip_when_match() throws Exception {
    Page page = new Page();
    page.setRawText("<div foo='foo'></div>");
    page.setRequest(new Request("http://codecraft.us/foo"));
    page.setUrl(PlainText.create("http://codecraft.us/foo"));
    ModelPageProcessor modelPageProcessor = ModelPageProcessor.create(null, ModelFoo.class, ModelBar.class);
    modelPageProcessor.process(page);
    assertThat(page.getResultItems().isSkip()).isFalse();
}
Example 9
Source File: MockGithubDownloader.java    From webmagic with Apache License 2.0 5 votes vote down vote up 
@Override
public Page download(Request request, Task task) {
    Page page = new Page();
    page.setRawText(html);
    page.setStatusCode(200);
    page.setRequest(new Request("https://github.com/code4craft/webmagic"));
    page.setUrl(new PlainText("https://github.com/code4craft/webmagic"));
    return page;
}
Example 10
Source File: CommonUtil.java    From blog-hunter with MIT License 4 votes vote down vote up 
private static Html getHtml(String html) {
    Page page = new Page();
    page.setRequest(new Request(""));
    page.setRawText(html);
    return page.getHtml();
}