public void runTestSimplePage() { // Page page = new Page(); page.urlStr = "http://www.google.de"; const String ankerUrlStr = "http://www.google.de/some_sub_path/"; const String imageUrlStr = "/bilder/bild1.jpg"; String imageUrlStrAbsolut = page.urlStr + imageUrlStr; String ankerTag = "<a \n href=\"" + ankerUrlStr + "\" attr2 = \"value2\">skjfk</a>"; String imageTag = "<img \r src\r\n=\n\r \"" + imageUrlStr + "\" some=\"other\" flags=\"flag\">"; page.pageContent = "jsdlkfjösjf ksjfj jdsflkjas f " + ankerTag + "sjlkfjsf" + imageTag + "sjdfkkdsjflk"; // CrawlerJobPartAnalyzePage crawlerJobPartAnalyzePage = new CrawlerJobPartAnalyzePage(); CrawlerJobContext crawlerJobContext = new CrawlerJobContext(); crawlerJobContext.page = page; crawlerJobContext.imageBacklog = new CrawlerImageBacklog(); crawlerJobPartAnalyzePage.crawlerJobContext = crawlerJobContext; // List <ICrawlerJobPart> result = crawlerJobPartAnalyzePage.run(); // Assert.IsNotNull(result); Assert.AreEqual(2, result.Count); Assert.IsTrue(result[0] is CrawlerJobPartGenerateSubJobs); Assert.IsTrue(result[1] is CrawlerJobPartRetrieveImages); // Assert.IsNotNull(page); Assert.IsNotNull(page.pageContent); Assert.IsNotNull(page.ankerHrefLinkList); Assert.AreEqual(1, page.ankerHrefLinkList.Count); Assert.AreEqual(ankerUrlStr, page.ankerHrefLinkList[0]); Assert.IsNotNull(page.imageSrcLinkList); Assert.AreEqual(1, page.imageSrcLinkList.Count); Assert.AreEqual(imageUrlStrAbsolut, page.imageSrcLinkList[0]); }
public void runTestMetataggedPage() { // Page page = new Page(); page.urlStr = "http://www.google.de"; const String ankerUrlStr = "http://www.google.de/some_sub_path/"; const String imageUrlStr = "/bilder/bild1.jpg"; String imageUrlStrAbsolut = page.urlStr + imageUrlStr; String ankerTag = "<a \n href=\"" + ankerUrlStr + "\" attr2 = \"value2\">skjfk</a>"; String imageTag = "<img \r src\r\n=\n\r \"" + imageUrlStr + "\" some=\"other\" flags=\"flag\">"; page.pageContent = "<mEta name=\"roBots\" content=\"ebc,noFollow,lala\">js<mEta name=\"roBots\" content=\"ebc,noIndEx,lala\">dlkfjösjf ksjfj jdsflkjas f " + ankerTag + "sjlkfjsf" + imageTag + "sjdfkkdsjflk"; // CrawlerJobPartAnalyzePage crawlerJobPartAnalyzePage = new CrawlerJobPartAnalyzePage(); CrawlerJobContext crawlerJobContext = new CrawlerJobContext(); crawlerJobContext.page = page; crawlerJobContext.imageBacklog = new CrawlerImageBacklog(); crawlerJobPartAnalyzePage.crawlerJobContext = crawlerJobContext; // List <ICrawlerJobPart> result = crawlerJobPartAnalyzePage.run(); // Assert.IsNotNull(result); Assert.AreEqual(0, result.Count); // Assert.IsNotNull(page); Assert.IsNotNull(page.pageContent); Assert.IsNull(page.ankerHrefLinkList); Assert.IsNull(page.imageSrcLinkList); }