/* Extract URLs (Web Addresses) From the Document */
// This script will scan all pages of the input document
// and extract :
// Valid URL s -- allows http, https, ftp, ftps and simple www.mydomain...
// Output PDF document will be placed in the same folder
// as input. The name of the output document will be:
// Original filename + "_Extracted_URLs"
// Visit www.evermap.com for more useful JavaScript samples.

var reMatch=/((((ht|f)tp(s?))\:\/\/)|(www.))([0-9a-zA-Z\-]+\.)+[a-zA-Z]{2,6}(\:[0-9]+)?(\/\S*)?/g;
//var reMatch=/(((ht|f)tp(s?))\:\/\/).*/g;

var strExt = "_Extracted_URLs.pdf";
var strIntro = "URLs (http,https,ftp,ftps) extracted from document: ";
var strFinal = "Total number of URLs (web addresses) extracted: " ;

ExtractFromDocument(reMatch,strExt,strIntro,strFinal);

function ExtractFromDocument(reMatch, strFileExt, strMessage1, strMessage2)
{
var chWord, numWords;

// construct filename for output document
var filename = this.path.replace(/\.pdf$/, strFileExt);

// create a report document
try {
    var ReportDoc = new Report();
    var Out = new Object(); // array where we will collect all our emails before outputing them
    
    ReportDoc.writeText(strMessage1 + this.path);
    ReportDoc.divide(1);      // draw a horizontal divider
    ReportDoc.writeText(" "); // write a blank line to output
    
    for (var i = 0; i < this.numPages; i++)
    {
        numWords = this.getPageNumWords(i);
        var PageText = "";
        for (var j = 0; j < numWords; j++) {
            var word = this.getPageNthWord(i,j,false);
            PageText += word;
            }

        var strMatches = PageText.match(reMatch);
        if (strMatches == null) continue;
        // now output matches into report document
        for (j = 0; j < strMatches.length; j++) {
            Out[strMatches[j]] = true; // store email as a property name
            }
    }
    
    var nTotal = 0;
    for (var prop in Out) 
    {
        ReportDoc.writeText(prop);
        nTotal++;
    }
    
    ReportDoc.writeText(" "); // output extra blank line
    ReportDoc.divide(1); // draw a horizontal divider
    ReportDoc.writeText(strMessage2 + nTotal);
    
    // save report to a document
    ReportDoc.save(
        {
        cDIPath: filename
        });

}
catch(e)
{
app.alert("Processing error: "+e)
}
    
} // end of the function