Coordinate Space Conversion
By specification, PDF documents have their own coordinate space, which is different from the coordinate space used by UIKit. While UIKit has its origin in the top-left corner, with y coordinates increasing in a downward direction, the y coordinates in PDF documents increase upward, starting in the bottom-left corner.
Additionally, the PDF coordinate space can be offset from the visible bottom-left corner of a page due to a crop box, and it can also be rotated due to the page being rotated. This is efficient because it means cropping or rotating a page is just setting a value: You don’t need to change the content stream or any of the annotations on the page. However, this can also be confusing to work with. Therefore, PSPDFKit exposes a normalized page coordinate space, which always puts the origin in the bottom-left corner of the visible area of the page.
Our Instant JSON format uses a coordinate space where the origin is the top-left corner of the page, with the y-axis increasing downward.
PSPDFKit provides an easy-to-use API for transforming between the PDF coordinate space and other coordinate spaces. For example, the following code uses the PSPDFKit conversion API to retrieve the onscreen coordinates of an annotation:
let pdfController: PSPDFViewController = ... let annotation: PSPDFAnnotation = ... if let pageView = pdfController.pageViewForPage(at: annotation.pageIndex) { // Get the (normalized) PDF coordinates of the annotation. let annotationPDFRect = annotation.boundingBox // Convert the annotation's PDF coordinates to view coordinates. let annotationViewRect = pageView.convert(rect: annotationPDFRect, from: pageView.pdfCoordinateSpace) }
PSPDFViewController *pdfController = ...; PSPDFAnnotation *annotation = ...; PSPDFPageView *pageView = [pdfController pageViewForPageAtIndex:annotation.pageIndex]; if (pageView) { // Get the (normalized) PDF coordinates of the annotation. CGRect annotationPDFRect = annotation.boundingBox; // Convert the annotation's PDF coordinates to view coordinates. CGRect annotationViewRect = [pageView convertRect:annotationPDFRect fromCoordinateSpace:pageView.pdfCoordinateSpace]; }
Understanding XFDF/PDF Rects
In addition to having their own coordinate spaces, PDFs also represent the bounding box of an annotation differently than PSPDFKit. For example, consider the case where we export the XFDF of a rectangle annotation like so:
<!-- Other attributes omitted for clarity --> <square rect="50.000000, 100.000000, 80.000000, 120.000000" />
The rect attribute contains the following information in this order:
-
The left side of the rectangle is 50 units from the left of the page.
-
The bottom side of the rectangle is 100 units from the bottom of the page.
-
The right side of the rectangle is 80 units from the left of the page.
-
The top side of the rectangle is 120 units from the bottom of the page.
The width of the rectangle annotation is 30 units (80-50) and the height is 20 units (120-100).
We already have the left value, but if we want to calculate the distance between the top side of the rectangle and the top of the page (PSPDFKit bounding boxes use width, height, left, and top values), we can subtract the XFDF top value from the height of the page. With a page height of 800, the adjusted top value would be 680 (800-120).
Thus, the equivalent PSPDFKit bounding box would be:
{
"top": 680,
"left": 50,
"width": 30,
"height": 20
}Conversion with Page Views
PSPDFPageView has a pdfCoordinateSpace property, which is an object conforming to the UICoordinateSpace protocol. This represents the coordinate space of that page view’s PDF page. Each UIView also conforms to UICoordinateSpace. Therefore, to convert between PDF page and view coordinates, you can call the following methods from UICoordinateSpace on either a PDF coordinate space or a view:
public func convert(_ point: CGPoint, to coordinateSpace: UICoordinateSpace) -> CGPoint public func convert(_ point: CGPoint, from coordinateSpace: UICoordinateSpace) -> CGPoint public func convert(_ rect: CGRect, to coordinateSpace: UICoordinateSpace) -> CGRect public func convert(_ rect: CGRect, from coordinateSpace: UICoordinateSpace) -> CGRect
- (CGPoint)convertPoint:(CGPoint)point toCoordinateSpace:(id <UICoordinateSpace>)coordinateSpace; - (CGPoint)convertPoint:(CGPoint)point fromCoordinateSpace:(id <UICoordinateSpace>)coordinateSpace; - (CGRect)convertRect:(CGRect)rect toCoordinateSpace:(id <UICoordinateSpace>)coordinateSpace; - (CGRect)convertRect:(CGRect)rect fromCoordinateSpace:(id <UICoordinateSpace>)coordinateSpace;
Page Info
PSPDFPageInfo can be used to get information for a specific page in a document. You can get this object by retrieving the PSPDFPageInfo object for a specific page on a PSPDFDocument using pageInfoForPageAtIndex:.
PSPDFPageInfo includes information like the size of a page, transforms, and page rotation.
Copying Annotations Across Rotated Pages
Note that some PDF files can also have rotation set on pages. You can get this value by checking the savedRotation property. This can be a value between 0 and 270 in 90-degree steps.
When a PDF page is rotated in the context of a PDF editor, it gets a rotation property assigned. However, its dimensions remain the same. See the attached screenshot showing the page information for the same page with different rotations applied.
Annotations in these rotated pages are positioned relative to the page dimensions. So if you, for example, copy an annotation on one page and then paste it onto the same page with an applied rotation, it’ll end up at the same position on the rotated page.
If the page is generated so it appears rotated without the rotation property applied (e.g. by generating a fresh PDF from a rotated image source), the copied annotation will no longer match the position and orientation in the original document. Instead of having a dimension of width × height with, say, 90 degrees of rotation, you now have a page of dimension height × width with 0-degree rotation. The annotations thus end up at the wrong positions in this new coordinate space.
The simplest solution is to ensure the coordinate spaces match. That means that rather than rotating the image, you create a PDF first and then rotate the page properly with a PDF editor or PDF library like PSPDFKit to preserve the rotation information. Then, copying annotations will just work.
If that’s not an option, you’ll need to update the annotation bounding boxes (and potentially even the annotation contents) to compensate for the switch in coordinates. PSPDFKit doesn’t provide a helper that would do that for you, because we don’t know that it’s the same page, just with the source content rotated. The bounding box rects need to be transformed and then applied to annotations. Using setBoundingBox:transform:includeOptional: might be useful to update the contents of the annotation if that’s also needed (passing transform: YES).
Conversion with Other Views
There are also conversion functions that can be used with any view. There is rarely a case where you need to use these functions, as the above property on PSPDFPageView is better suited for most use cases. Use these functions only when the page is shown in a view other than a PSPDFPageView. For these to work correctly, the page must fill the view’s bounds:
func PSPDFConvertViewPointToPDFPoint(_ viewPoint: CGPoint, _ pageInfo: PSPDFPageInfo, _ viewBounds: CGRect) -> CGPoint func PSPDFConvertPDFPointToViewPoint(_ pdfPoint: CGPoint, _ pageInfo: PSPDFPageInfo, _ viewBounds: CGRect) -> CGPoint func PSPDFConvertPDFRectToViewRect(_ pdfRect: CGRect, _ pageInfo: PSPDFPageInfo, _ viewBounds: CGRect) -> CGRect func PSPDFConvertViewRectToPDFRect(_ viewRect: CGRect, _ pageInfo: PSPDFPageInfo, _ viewBounds: CGRect) -> CGRect
CGPoint PSPDFConvertViewPointToPDFPoint(CGPoint viewPoint, PSPDFPageInfo *pageInfo, CGRect viewBounds); CGPoint PSPDFConvertPDFPointToViewPoint(CGPoint pdfPoint, PSPDFPageInfo *pageInfo, CGRect viewBounds); CGRect PSPDFConvertPDFRectToViewRect(CGRect pdfRect, PSPDFPageInfo *pageInfo, CGRect viewBounds); CGRect PSPDFConvertViewRectToPDFRect(CGRect viewRect, PSPDFPageInfo *pageInfo, CGRect viewBounds);
To learn how to convert points, lines, and rectangles from view coordinates to PDF coordinates, refer to the next sections.
Converting Points
let annotation: Annotation = ... let pageIndex = ... let pageInfo = document.pageInfoForPage(at: pageIndex)! let view: UIView = ... // Some `UIView` that displays a PDF page filling the view. let viewBounds = view.bounds let viewPoint = CGPoint(x: 100, y: 200) // Convert view points into PDF points. let pdfPoint = PSPDFConvertViewPointToPDFPoint(viewPoint, pageInfo, viewBounds) annotation.points = [pdfPoint]
PSPDFAnnotation *annotation = ... PSPDFPageIndex pageIndex = ... PSPDFPageInfo *pageInfo = [document pageInfoForPageAtIndex:pageIndex]; UIView *view = ... // Some `UIView` that displays a PDF page filling the view. CGRect viewBounds = view.bounds; CGPoint viewPoint = CGPointMake(100, 200); // Convert view points into PDF points. CGPoint pdfPoint = PSPDFConvertViewPointToPDFPoint(viewPoint, pageInfo, viewBounds); annotation.points = @[@(pdfPoint)];
Converting Lines
let inkAnnotation: InkAnnotation = ... let pageIndex = ... let pageInfo = document.pageInfoForPage(at: pageIndex)! let view: UIView = ... // Some `UIView` that displays a PDF page filling the view. let viewBounds = view.bounds let viewLines = [ [(CGPoint(x: 100, y: 100)), (CGPoint(x: 100, y: 200)), (CGPoint(x: 150, y: 300))], [(CGPoint(x: 200, y: 100)), (CGPoint(x: 200, y: 200)), (CGPoint(x: 250, y: 300))] ] // Convert view line points into PDF line points. let pdfLines = ConvertToPDFLines(viewLines: viewLines, pageInfo: pageInfo, viewBounds: viewBounds) inkAnnotation.lines = pdfLines
PSPDFInkAnnotation *inkAnnotation = ... PSPDFPageIndex pageIndex = ... PSPDFPageInfo *pageInfo = [document pageInfoForPageAtIndex:pageIndex]; UIView *view = ... // Some `UIView` that displays a PDF page filling the view. CGRect viewBounds = view.bounds; NSArray *viewLines = @[ @[@(CGPointMake(100, 100)), @(CGPointMake(100, 200)), @(CGPointMake(150, 300))], @[@(CGPointMake(200, 100)), @(CGPointMake(200, 200)), @(CGPointMake(250, 300))] ]; // Convert view line points into PDF line points. NSArray *pdfLines = PSPDFConvertViewLinesToPDFLines(viewLines, pageInfo, viewBounds); inkAnnotation.lines = pdfLines;
Converting Rectangles
let annotation: Annotation = ... let pageIndex = ... let pageInfo = document.pageInfoForPage(at: pageIndex)! let view: UIView = ... // Some `UIView` that displays a PDF page filling the view. let viewBounds = view.bounds let viewRect = CGRect(x: 100, y: 100, width: 300, height: 300) // Convert view rects into PDF rects. let pdfRect = PSPDFConvertViewRectToPDFRect(viewRect, pageInfo, viewBounds) annotation.rects = [pdfRect]
PSPDFAnnotation *annotation = ... PSPDFPageIndex pageIndex = ... PSPDFPageInfo *pageInfo = [document pageInfoForPageAtIndex:pageIndex]; UIView *view = ... // Some `UIView` that displays a PDF page filling the view. CGRect viewBounds = view.bounds; CGRect viewRect = CGRectMake(100, 100, 300, 300); // Convert view rects into PDF rects. CGRect pdfRect = PSPDFConvertViewRectToPDFRect(viewRect, pageInfo, viewBounds); annotation.rects = @[@(pdfRect)];
Converting between the Normalized PDF Coordinate Space and the Raw PDF Coordinate Space
The PSPDFKit API uses a normalized page coordinate space, which always puts the origin in the bottom-left corner of the visible area of the page.
You can convert from the raw PDF coordinate space to PSPDFKit’s normalized space using the transform property of PSPDFPageInfo. For example, to place a 100-point by 100-point square annotation at the page origin stored in the PDF, use code like the following:
let document: PSPDFDocument = ... let pageInfo = document.pageInfoForPage(at: 13) let rawPDFRect = CGRect(x: 0, y: 0, width: 100, height: 100) let normalizedPDFRect = rawPDFRect.applying(pageInfo.transform) let annotation: PSPDFSquareAnnotation = ... annotation.boundingBox = normalizedPDFRect
PSPDFDocument *document = ... PSPDFPageInfo *pageInfo = [document pageInfoForPageAtIndex:13]; CGRect rawPDFRect = CGRectMake(0, 0, 100, 100); CGRect normalizedPDFRect = CGRectApplyAffineTransform(rawPDFRect, pageInfo.transform); PSPDFSquareAnnotation *annotation = ... annotation.boundingBox = normalizedPDFRect;
The annotation might not end up visible if there is a crop box offset.
To convert from the normalized PDF coordinate space used by PSPDFKit to the raw PDF coordinate space, use the inverse of the page info’s transform. For example, to read an annotation’s bounding box as it would be serialized in the PDF, use the following:
let document: PSPDFDocument = ... let pageInfo = document.pageInfoForPage(at: 13) let annotation: PSPDFAnnotation = ... let normalizedPDFRect = annotation.boundingBox let rawPDFRect = normalizedPDFRect.applying(pageInfo.transform.inverted())
PSPDFDocument *document = ...
PSPDFPageInfo *pageInfo = [document pageInfoForPageAtIndex:13];
PSPDFAnnotation *annotation = ...
CGRect normalizedPDFRect = annotation.boundingBox;
CGRect rawPDFRect = CGRectApplyAffineTransform(normalizedPDFRect, CGAffineTransformInvert(pageInfo.transform));How to Convert between Raster Image Pixels and Points
The concept of resolution does not apply to PDF documents unless they have been converted into raster images, i.e. images whose dimensions are expressed in pixels. The default unit PSPDFKit returns for page sizes is the point, which is easily converted into inches by considering the fact that 1 inch is equal to 72 points. Inch separation results from dividing the size in points of a particular page by 72. Resolution, expressed in DPI (dots per inch), is thus the result of dividing the page size in points by the inch separation. To summarize, these are the relations you have to consider when you need to convert between points/inches and pixels in PDF:
1 inch = 72 points
Inch separation = points / 72
DPI (resolution) = pixels / inch separation
Note that, since PDF 1.6, the relationship between inches and points may be specified as greater than 1⁄72 by means of the UserUnit entry of the page dictionary. See table 30 on page 79 of the PDF 1.7 specification for more information.