Some documents have a vertical text in the margin which is needed for identification in subsequent steps of the post-processing. Also, table headers with vertical text are used occasionally.

To test text which is rotated by an arbitrary angle, the same test methods are available as for “normal” text.

The document used by the next example contains two vertical texts:

And this is the test using a page region for the vertical text:

// Comparing upright text in a part of a PDF page
@Test
public void hasRotatedTextInRegion() throws Exception {
  String filename = "verticalText.pdf";
  
  int leftX  =  40;
  int upperY =  45;
  int width  =  45;
  int height = 110;
  PageRegion pageRegion = new PageRegion(leftX, upperY, width, height);

  String textBottomToTop = "Text from bottom to top.";
  String textTopToBottom = "Text from top to bottom.";
  AssertThat.document(filename)
            .restrictedTo(FIRST_PAGE)
            .restrictedTo(pageRegion)
            .hasText()
            .containing(textBottomToTop, WhitespaceProcessing.IGNORE)
            .containing(textTopToBottom, WhitespaceProcessing.IGNORE)
  ;
}

The parser extracts vertical text using a lot of line breaks. Therefore, it is absolutely necessarry to use WhitespaceProcessing.IGNORE.

Note that the vertical text is still text with the writing direction LTR (left-to-right). PDFUnit also supports RTL (right-to-left) text; see chapter 3.34: “Text - Right to Left (RTL)”.