public void TestWord95() { // Too old for the default try { extractor = new WordExtractor( POIDataSamples.GetDocumentInstance().OpenResourceAsStream("Word95.doc") ); Assert.Fail(); } catch (OldWordFileFormatException) { } // Can work with the special one Word6Extractor w6e = new Word6Extractor( POIDataSamples.GetDocumentInstance().OpenResourceAsStream("Word95.doc") ); String text = w6e.Text; Assert.IsTrue(text.Contains("The quick brown fox jumps over the lazy dog")); Assert.IsTrue(text.Contains("Paragraph 2")); Assert.IsTrue(text.Contains("Paragraph 3. Has some RED text and some BLUE BOLD text in it")); Assert.IsTrue(text.Contains("Last (4th) paragraph")); String[] tp = w6e.ParagraphText; Assert.AreEqual(7, tp.Length); Assert.AreEqual("The quick brown fox jumps over the lazy dog\r\n", tp[0]); Assert.AreEqual("\r\n", tp[1]); Assert.AreEqual("Paragraph 2\r\n", tp[2]); Assert.AreEqual("\r\n", tp[3]); Assert.AreEqual("Paragraph 3. Has some RED text and some BLUE BOLD text in it.\r\n", tp[4]); Assert.AreEqual("\r\n", tp[5]); Assert.AreEqual("Last (4th) paragraph.\r\n", tp[6]); }
public void TestWord6() { // Too old for the default try { extractor = new WordExtractor( POIDataSamples.GetDocumentInstance().OpenResourceAsStream("Word6.doc") ); Assert.Fail(); } catch (OldWordFileFormatException) { } Word6Extractor w6e = new Word6Extractor( POIDataSamples.GetDocumentInstance().OpenResourceAsStream("Word6.doc") ); String text = w6e.Text; Assert.IsTrue(text.Contains("The quick brown fox jumps over the lazy dog")); String[] tp = w6e.ParagraphText; Assert.AreEqual(1, tp.Length); Assert.AreEqual("The quick brown fox jumps over the lazy dog\r\n", tp[0]); }