The Situation: I have a pdf document that includes hyperlinks. I would like my iPhone app to display this pdf document, and enable clicking of the hyperlinks.
Note: Currently, I am trying to do this with a UIWebView, which unfortunately seems unable to detect link taps. There is a family of classes & methods called CGPDF which I have read about in the documentation. The documentation talks all about parsing pdf documents and what not but is not specific on how to detect link taps and what not (thanks apple).
I can't comment on UIWebView, but the CGPDF stuff is part of CoreGraphics. So it's primarily for drawing — you'll have a PDF object and a context to draw it to (almost certainly received via a custom subclass of UIView) and ask it to draw at a certain scale and position. If you're new to CoreGraphics and/or custom UIView subclasses then it might take a day or so to get into it, but it'll end up being a few hundred lines at worst.
CoreGraphics doesn't know anything about user interaction, so I think Apple considered it implicit that there's no concept of taps down there, and hence no way that CoreGraphics can directly help you with catching taps.
There's example code for loading a PDF and drawing it to a context at the top of this document. CoreGraphics is the same on OS X and iOS, except that the initial coordinate system is the opposite way up. On OS X it is the same way around as a PDF, on iOS it's the opposite way, so PDFs will render inverted along y if otherwise unadjusted. The section directly after the example display code is entitled 'Creating a Transform for a PDF Page' and should help with that — you just need to apply a transform that inverts y.
There's a tutorial here on how to create a custom UIView subclass that includes some CoreGraphics drawing steps.
With respect to catching taps, things get a bit more complicated. You don't need to do anything like a full PDF inspection because of the way the file format works. Things that are links are drawn to look however they should look per the normal PDF operations that CoreGraphics can handle for you. There are then separate tables that describe bounding box rectangles for catching taps and the links they should go to. Those tables are parsed into an easy to read format by CoreGraphics but not inspected. So you're given all the tools to find out where the links are without having to learn too much about the file format, but you still need to fish through for them yourself.
The relevant CoreGraphics construct is the CGPDFDictionary. You can grab the latest PDF specification here, but 99% of it isn't relevant to you or covers things that CoreGraphics already implements. You'll need to jump back and forth through the documentation to piece all the bits together, but you can leap in at section 8.4 on Annotations, on Page 488.
It ends up being a reasonably complicated combination of bits of code and I can't post my own for commercial reasons but you'll want to build up a mapping of page names to page numbers, traverse the link annotations (which are links to named pages) and do something to catch taps. Personally I just added invisible UIButtons as subviews of my custom PDF view. Obviously any transform you're applying to the PDF needs to be applied to the annotation coordinates too.
The total thing was maybe 1,000 lines.
Related
I need access to the coordinates of individual strokes of a PKDrawing in PencilKit. Is there any way I can get access to that? Currently, my only idea is to try and decode the opaque data representation we get from PKDrawing.
As Ben has said, there doesn't seem to be any way to access stroke-level data in PencilKit at this time. This seems like a pretty rudimentary feature so hopefully Apple will add it next WWDC. Fingers crossed.
LetsBuildThatApp's YouTube tutorial is a good starting point if you just want basic drawing capabilities and are not too concerned about drawing quality or latency. I ran into issues when I tried to add the ability to vary the stroke width with pen pressure. I was never able to get it to transition smoothly between stroke segments of different widths - it always 'jumped' jarringly from one width to the next. Maybe there's a way to fix that, but I wasn't able to.
I suspect that currently, the only way to draw low-latency high-quality images with Apple Pencil is to write a drawing engine from scratch in Metal or OpenCL. It's strange it has taken Apple so long since the release of Apple Pencil to get a comprehensive drawing framework out. Fingers crossed that changes at WWDC 2020.
Edit: Apple announced substantial updates to PencilKit at WWDC20, including access to stroke data and functions to programatically draw strokes.
I was was looking for the same thing and from this comment https://stackoverflow.com/a/57565661/8891611 and my own searching of the documentation, it seems what you're asking for doesn't exist. If you do discover a way to decode the opaque data please update this thread on how so.
If you aren't necessarily set on PencilKit, you could manually code your drawing functions which is easier than it sounds. In this tutorial: https://youtu.be/E2NTCmEsdSE?t=504 I have timestamped where he shows that he has access to all the points on the lines he is creating.
And another potential solution could be to use both, by have a mapping between your PencilKit drawing and the points found from the technique in this tutorial.
For the strokes, you can access the property .renderBounds that gives you the dimensions of the rectangle where the stroke is into.
All you ios architects out there, please help me choose architecture/technology for the following iphone/ipad app.
The app itself is a financial app, but we want more of a game look-and-feel of the app, so we probably don't want to use the builtin looks of the cocoa widgets. The elements on the screen will probably be some kind of blob-shaped images.
The app will essentially have five "blob"-shaped areas, spread out evenly across the screen. One of the blobs will be centered and larger than the other ones. Within each blob there will be clickable areas which will pop up "details" and menu-action blobs. These blobs are also graphics objects and must not take over the whole screen. The blobs should animate nicely when popping up. The graphics elements will have a couple of lines of text, which are generated, so the overlaying text itself cannot be part of the static background-image.
The main user interaction will be swiping within the center blob, displaying summaries of the items that are conceptually contained within the blobs underlying data store. Now and then, the user will drag and drop the item to one of the other blobs. While dragging, the item should be traced by a line and when dropping on the other blob, the item should be animated to look like it's being "sucked into" the blob.
Now, what kind of technique would you suggest for this app? Is Cocoa suitable in this scenario? Should I use a game framework like Cocos2D? All kinds of suggestions including example code snippets are most welcome.
I realize that this question might not be as straightforward and to the point as questions generally are on SO, but I hope your answers will come to use by more people than me. Thanks!
EDIT (MY SOLUTION):
I eventually ended up doing everything in UIKit, which was a lot easier than I expected.
Briefly described I used UIButtons with Custom style and an image background, which gave me full control over the visual appearance of the "items". I also found it very useful to manipulate the underlying CALayer of many of my other UIViews. It is often easier than drawing things from scratch using Core Graphics programming.
Another thing that was useful were the UIGestureRecognizer:s. I found them useful for both handling "real" gestures like swiping, longpress etc, but also for handling normal "tap" for UIView classes that aren't subclasses of UIControl. Two examples are UIImage, UILabel and UIView itself. That way I could handle taps for these simple classes. I could for example use a normal UIView, modify it's CALayer to change the look of it completely and still handle taps. Using this technique, I didn't have to subclass any views at all in my app.
The animations were pretty easy too, even though I had to use a non-public method to use "suck" animation, so my app will never pass App Store moderation. It was just a prototype anyway so I don't care.
When this app will be for real, I will probably implement it in HTML5/JavaScript wrapped by Phonegap. The reason for this is mainly reuse of existing mobile web services and also for code reuse across platforms. It will probably also be easier to hook into the existing security solution when using a webapp.
Cocos2d is great if you need to move elements around really fast as it is a layer on top of OpenGLES. I think from what you have said the UIKit will be fine, you get nice animation support, you can do some nice things with UIScrollViews to handle moving elements around etc.
If you need more detailed graphics support and lots of moving elements, particle effects etc then by all means go for Cocos2D but be aware that in Cocos2d the application works more on a scheduled update method, i.e. you get notified every 1/60th of a second to move stuff draw stuff etc, whereas with normal UIKit approach it is more event drive, i.e. I click a button and show a view etc.
I'm pretty new to iphone development, so this is more of a high-level question. The simplest description of what I am looking to do is create a zoomable/panable field on which I can place a bunch of circle objects. The number of these circles is likely to be in the hundreds, and ideally when the user zooms in close enough, more information can be displayed. From stuff I've read, it seems like UIScrollView provides the simplest way of making a zoomable/panable view but I'm not sure it's the best way to handle a view that includes a hundred graphic objects. I'm trying to figure out if I should progress further down that path or look into things like CALayers, Core Graphics, etc. Any guidance or advice would be greatly appreciated. Thanks in advance,
Roman
I suggest you to use UIScrollView, because it will save a lot of time for handling proper zooming/scrolling. So the workflow is next:
1. Zoom you scroll view
2. In delegate's callback scrollViewDidEndZooming:withView:atScale: you can obtain the scale and determine the level of detail that you need.
3. redraw the visible region (using Core Graphics) with appropriate level of detail (number of circles etc.)
So you should use the mix of Core Graphics and UIKit.
I am creating a syntax highlighter for the iPhone and in order to display text with multiple formats, I have sub-classed UIView and modified the drawRect: method so that each line is displayed with the proper syntax highlighting (highlighting is done earlier with RegEx, text is drawn with CGContextShowTextAtPoint() one line at a time). Everything works ok, I store each line of text as an NSString in an NSMutableArray, I handle the keyboard through a hidden UITextField and its delegate methods, the cursor is a blinking CALayer that can be moved around with touches and I have a custom scroll view that handles scrolling. However, I have two problems that I can't seem to wrap my head around:
Word wrap, right now the text just keeps going off the left end of the screen. To keep things fast I only redraw the portions of the view that have changed (usually just the line being edited, but sometimes the lines below as well e.g. if you press return halfway through the document) with setNeedsDisplayInRect:. This makes word wrap complicated because then you have to draw more than one line on the screen, even though it still is only one object in the array.
UIViews have a maximum content size of 1024x1024 which equates to about 64 lines. I need the ability to display more than that. I am thinking about using multiple CALayers one after another, but I am having trouble drawing content to the layers (using drawLayer:inContext: and drawInContext:).
So my questions are:
Does anyone have any, even general, suggestions about how to accomplish either of these two points. Or,
Has someone already written a custom text-drawing view that handles these things that I could use instead.
Thanks,
Kyle
EDIT: The scrolling problem is pretty much solved, however I am still having trouble with word-wrap. My trouble is that everything is done by line: the view updates one line at a time, the text is stored as an array of lines, the highlighter highlights one line at a time, etc. and having a single index in the array (one line of text) take up multiple lines on the screen raises some problems, for example, I had to implement my own movable cursor and when you move the cursor it needs to be able to turn a display line (found by dividing touch.x by the line height) into a text line (an index in the array). Any ideas?
You should first spend some quality time understanding how this problem was solved on Mac:
http://developer.apple.com/documentation/Cocoa/Conceptual/TextArchitecture/Tasks/AssembleSysByHand.html
http://developer.apple.com/documentation/Cocoa/Conceptual/TextLayout/TextLayout.html
In particular, you should become familiar with line fragment generation, which is the problem you're trying to solve for word-wrap, and you should understand glyph generation in order to do rich text well. You don't need to implement all of the flexibility of NSTypesetter or NSLayoutManager of course, but the text system on Mac is incredibly powerful, and you should learn from its example. Implementing something similar to NSAttributedString for iPhone may be valuable to you and improve your performance.
Unless you're moving things around a lot (for which this would best work in a UIScrollView), I don't think you should need to use CALayers here. That seems overkill for the problem, and may actually adversely impact the optimizations already provided by UIScrollView. If you're seeing performance problems, first make sure you're not doing redundant calculations within your drawRect:.
Check out TTStyledText in Three20 library. Not sure how well it matches your goals, but might serve you as an example. (The library itself is a bit bloated, but is a wonderful source to look at.)
It may be best to draw your text within a CATiledLayer hosted within your UIView, in order to get around the 1024x1024 texture size limit (which appears to actually be 2048x2048 on all existing devices). For an example of text drawing within a CALayer, I'd refer you to the CPTextLayer class within the Core Plot framework. This layer (which inherits from the CPLayer class within that same framework) does cross-platform (Mac and iPhone) text rendering in a CALayer. You might be able to extend it to work as a CATiledLayer for longer text blocks.
One thing to be aware of is that platform-specific drawAtPoint: methods are used in this layer, instead of the CGContextShowTextAtPoint() function you are using. The reason for this is that CGContextShowTextAtPoint() only works on ASCII text, meaning that you can't do Unicode text rendering with it. There is an example of using CGContextShowTextAtPoint() to draw text, within a #define'd out portion at the bottom of the renderAsVectorInContext: method.
This might be trivial for some of you, but I have two screenshots from the Lose It! app in which I'm curious how two different screens were put together.
The first:
That middle graph which shows the statistics chart. Is that a custom image being drawn on top of with Core Graphics / Quartz to achieve the desired numbers? Is the yellow line that's being dynamically allocated all the work of Quartz?
And second:
This one might be a bit easier, but the whole bar which looks like a native UIKit widget, which contains [Budget, Food, Exercise, Net, Under]. There appears to be a drop shadow above it. Are they doing a drop shadow on the UINavigationBar? Is the menu below it just a UIImage that a designer was able to craft to look like the UINavigationBar?
If there's a blog out there which teaches UI tricks such as these, I'd love to read more.
1) Yes, it's likely a view that uses the chart as a background and then uses core graphics to render the line,
2) This could be a single view divided into four sections. Each section has two lines of text drawn with different colors. It's possible that each section may be a view that encapsulates this behavior.
I'm not aware of any blog that teaches these "tricks". It's really a case of understanding what functionality is available and then using it creatively to develop your UI.
For example we know it's possible to;
Draw images at different sizes/positions.
Draw text in different fonts, sizes, colors, alignment
Draw primitives
Really, when you have those you can create pretty much anything.
I think there's an SDK sample that demonstrates using custom views to create a fancy timezone style applications. That might be one worth checking out.
Update: found it, it's here.