How can I parse content from a PDF page with Swift - ios

Question

The documentation is not really clear to me. So far I reckon I need to set up a CGPDFOperatorTable and then create a CGPDFContentStreamCreateWithPage and CGPDFScannerCreate per PDF page.
The documentation refers to setting up Callbacks, but it's unclear to me how. How to actually obtain the content from a page?
This is my code so far.
let pdfURL = NSBundle.mainBundle().URLForResource("titleofdocument", withExtension: "pdf")
// Create pdf document
let pdfDoc = CGPDFDocumentCreateWithURL(pdfURL)
// Nr of pages in this PF
let numberOfPages = CGPDFDocumentGetNumberOfPages(pdfDoc) as Int
if numberOfPages <= 0 {
// The number of pages is zero
return
}
let myTable = CGPDFOperatorTableCreate()
// lets go through every page
for pageNr in 1...numberOfPages {
let thisPage = CGPDFDocumentGetPage(pdfDoc, pageNr)
let myContentStream = CGPDFContentStreamCreateWithPage(thisPage)
let myScanner = CGPDFScannerCreate(myContentStream, myTable, nil)
CGPDFScannerScan(myScanner)
// Search for Content here?
// ??
CGPDFScannerRelease(myScanner)
CGPDFContentStreamRelease(myContentStream)
}
// Release Table
CGPDFOperatorTableRelease(myTable)
It's a similar question to: PDF Parsing with SWIFT but has no answers yet.

Answer 1

Here is an example of the callbacks implemented in Swift:
let operatorTableRef = CGPDFOperatorTableCreate()
CGPDFOperatorTableSetCallback(operatorTableRef, "BT") { (scanner, info) in
print("Begin text object")
}
CGPDFOperatorTableSetCallback(operatorTableRef, "ET") { (scanner, info) in
print("End text object")
}
CGPDFOperatorTableSetCallback(operatorTableRef, "Tf") { (scanner, info) in
print("Select font")
}
CGPDFOperatorTableSetCallback(operatorTableRef, "Tj") { (scanner, info) in
print("Show text")
}
CGPDFOperatorTableSetCallback(operatorTableRef, "TJ") { (scanner, info) in
print("Show text, allowing individual glyph positioning")
}
let numPages = CGPDFDocumentGetNumberOfPages(pdfDocument)
for pageNum in 1...numPages {
let page = CGPDFDocumentGetPage(pdfDocument, pageNum)
let stream = CGPDFContentStreamCreateWithPage(page)
let scanner = CGPDFScannerCreate(stream, operatorTableRef, nil)
CGPDFScannerScan(scanner)
CGPDFScannerRelease(scanner)
CGPDFContentStreamRelease(stream)
}

Answer 2

You've actually specified exactly how to do it, all you need to do is put it together and try until it works.
First of all, you need to setup a a table with callbacks as you state yourself in the beginning of your question (all code in Objective C, NOT Swift):
CGPDFOperatorTableRef operatorTable = CGPDFOperatorTableCreate();
CGPDFOperatorTableSetCallback(operatorTable, "q", &op_q);
CGPDFOperatorTableSetCallback(operatorTable, "Q", &op_Q);
This table contains a list of the PDF operators you want to get called for and associates a callback with them. Those callbacks are simply functions you define elsewhere:
static void op_q(CGPDFScannerRef s, void *info) {
// Do whatever you have to do in here
// info is whatever you passed to CGPDFScannerCreate
}
static void op_Q(CGPDFScannerRef s, void *info) {
// Do whatever you have to do in here
// info is whatever you passed to CGPDFScannerCreate
}
And then you create the scanner and get it going, while passing it the information you just defined.
// Passing "self" is just an example, you can pass whatever you want and it will be provided to your callback whenever it is called by the scanner.
CGPDFScannerRef contentStreamScanner = CGPDFScannerCreate(contentStream, operatorTable, self);
CGPDFScannerScan(contentStreamScanner);
If you want to see a complete example with sourcecode on how to find and process images, check this website.

Answer 3

To understand why a parser works this way, you need to read the PDF specification a bit better. A PDF file contains something close to printing instructions. Such as "move to this coordinate, print this character, move there, change the color, print the character number 23 from the font #23", etc.
The parser gives you callbacks for each instructions, with the possibility to retrieve the instruction parameters. That's all.
So, in order to get the content from a file, you need to rebuild its state manually. Which means, recompute the frames for all characters, and try to reverse-engineer the page layout. This is clearly not an easy task, and that's why people have created libraries to do so.
You may want to have a look at PDFKitten , or PDFParser which is a Swift port with some improvement that i did.