You are right, there are no samples in the SDK to demonstrate this. The reason is that this is not a stright forward process. There are two appraoches you can take: You iterate over all words in the document and get the "Quads" information, then you need to assemble these elements in the right order and figure out what information you are looking for based on the location and size you can extract from the quads. The other way is to crop down the page to just the area you are interested in, get all the words in that area (and potentially the quads again) and extract it this way.
The functions you are interested in are Doc.getPageNumWords(), Doc.getPageNthWord() and Doc.getPageNthWordQuads()
This does not always work as you want it to work however... The results depend highly on the quality of your PDF file, and the kind of information you need to extract. One problem is that you only get "words", no puctuation marks.