User:Pedro Sá Couto/Prototyping 3rd

Watermarks, downloading, deleting

https://pad.xpub.nl/p/IFL_2018-05-13

JSOR (Ithaka Harbors, Inc) publishers

• download an article from http://jstor.org/ from within the school´s internet
• search for watermarks on the PDF
• look for some leads in https://sourceforge.net/p/pdfedit/mailman/message/27874955/ — https://github.com/kanzure/pdfparanoia/issues?utf8=%E2%9C%93&q=jstor

Gather Direct messages

1. While working around the watermarks that JSTOR leaves in their PDFs, we understood that in their website, one can have a preview through the whole document without any of these watermarks. The idea began to download it directly from these preview images rather than actually downloading it from a JSTOR account.
2. https://www.jstor.org/stable/23267102?Search=yes&resultItemClick=true&searchText=tea&searchUri=%2Faction%2FdoBasicSearch%3FQuery%3Dtea%26amp%3Bacc%3Don%26amp%3Bfc%3Doff%26amp%3Bwc%3Don%26amp%3Bgroup%3Dnone&ab_segments=0%2Fl2b-basic-1%2Frelevance_config_with_tbsub_l2b&refreqid=search%3A9fd0deff8d3258de87d3b54d6dfad664&seq=1#metadata_info_tab_contents the idea is to create a script that iterates through the url sequence number("seq=i")
3. https://git.xpub.nl/pedrosaclout/jsort_scrape

#SCRIPT GOES HERE WHEN FINISHED

• 
• 

Verso Books

• Create an account in https://www.versobooks.com/
• Download book as epub, a free one will do, for instance https://www.versobooks.com/books/2772-verso-2017-mixtape
• search for watermarks on the EPUB
• Look at the Institute for Biblio-Immunology -- First Communique: https://pastebin.com/raw/E1xgCUmb for more leads. https://www.booxtream.com/

• 
• 

Tesseract, OCR, Book scan

<script type="text/javascript">

    //store all class 'ocr_line' in 'lines'
    var lines = document.querySelectorAll(".ocr_line");    

    //loop through each element in 'lines'
    for (var i = 0; i < lines.length; i++){ 

      var line = lines[i];
      console.log(line.title) 

      //split the content of 'title' every space and store the list in 'parts'
      var parts = line.title.split(" ");
      console.log(parts);

      // width and height starts from the side 
      var left = parseInt(parts[1], 10);
      var top = parseInt(parts[2], 10);
      var width = (parseInt(parts[3], 10) - left);
      var height = (parseInt(parts[4], 10) - top);

      // create a style element with the content selected from the list 'parts'
      line.style = "position: absolute; left: " + parts[1] + "px; top: " + parts[2] + "px; width: " + width + "px; height: " + height + "px; border: 5px solid lightblue";

      var words = line.querySelectorAll(".ocrx_word");

      for (var e = 0; e < words.length; e++){ 

        var span = words[e];
        console.log(span.title) 

        var parts = span.title.split(" ");
        console.log(parts);

        var wleft = parseInt(parts[1], 10);
        var wtop = parseInt(parts[2], 10);
        var wwidth = (parseInt(parts[3], 10) - wleft);
        var wheight = (parseInt(parts[4], 10) - wtop);

        span.style = "position: absolute; left: " + (wleft - left) + "px; top: " + (wtop - top) + "px; width: " + wwidth  + "px; height: " + wheight + "px; border: 2px solid purple";
      } 
    }

• 
• 
• 
•