Image Cookbook

Get all document images

const doc = new npdf.Document()
doc.load('/path/to/document.pdf', (err, data) => {
  if(e) {/* handle error */}
  // iterate the document body, filter type XObject with subtype Image
  // iterating the body can be expensive in terms of javascript / c++ copy overhead
  for(let i = 0; i < doc.body.length; i++) {
    const o = doc.body[i]
    if (o.type === 'Dictionary') {
        let objDict = o.getDictionary(),
            objType = objDict.hasKey('Type') ? objDict.getKeyType('Type') : null,
            objSubType = objDict.hasKey('SubType') ? objDict.getKeyType('SubType') : null

        if ((objType && objType === 'Name') ||
            (objSubType && objSubType === 'Name')) {

            if ((objType && objDict.getKey<String>('Type') === 'XObject') || (objSubType && objDict.getKey<String>('SubType') === 'Image')) {
                if (objDict.hasKey('Filter')) {
                    if (objDict.getKeyType('Filter') === 'Array') {
                        const imgObjArr = objDict.getKey<npdf.Array>('Filter')
                        if (imgObjArr.length === 1) {
                            if ((imgObjArr.at(0) as npdf.Object).type === 'Name') {
                                if ((imgObjArr.at(0) as npdf.Object).getName() === 'DCTDecode') {
                                    extractImg(o, true)
                                    return
                                }
                            }
                        }
                    } else if (objDict.getKeyType('Filter') === 'Name' && objDict.getKey<String>('Filter') === 'DCTDecode') {
                        extractImg(o, true)
                        return
                    }
                }
            }
        }
    }
  }
})
function extractImg(obj: npdf.Object, jpg: Boolean) {
    let ext = jpg ? '.jpg' : '.ppm'
    writeFile(`/tmp/${v4()}.${ext}`, obj.stream, err => {
        if (err instanceof Error) { /* handle error */}
    })
}

Get an image from a specific page

Getting images from a page is similar to the above with the exception that instead of iterating the document's body we will inspect a page's resources dictionary. This method will only work if the image has been set in the page's resources. If the image is an inline object the following method will not work, for finding inline images the above method must be used.

const doc = new npdf.Document()
doc.load('/path/to/doc', e => {
    if (e instanceof Error) t.fail()
    const page = doc.getPage(0)
    let resource = page.resources.getDictionary().getKey<npdf.Dictionary>('XObject')
    let xKey = resource.getKeys()
    let xob = resource.getKey<npdf.Dictionary>('XOb4')
    if(xob.getKey<String>(NPDFName.SUBTYPE) === 'Image') {
        if(xob.hasKey('Filter') && xob.getKey<String>('Filter') === 'DCTDecode') {
            extractImg(xob.obj, true)
        } else {
            extractImg(xob.obj, true)
        }
    }
})
function extractImg(obj: npdf.Object, jpg: Boolean) {
    // img data is accessable via npdf.Object.stream
}

Set an image

import {nopodofo as npdf} from 'nopodofo'
const doc = new npdf.StreamDocument()
const page = doc.createPage(new npdf.Rect(0, 0, 612, 792))
const painter = new npdf.Painter(doc)
const img = new npdf.Image(doc, join(__dirname, '../../spec/test-documents/test.jpg'))
painter.setPage(page)
painter.drawImage(img, 0, page.height - img.height)
painter.finishPage()