Question

我正在尝试使用PyObjC覆盖带有一些文本的图像，同时努力回答我的问题，"Annotate images using tools built into OS X"。通过引用CocoaMagic，RMagick的RubyObjC替换，我想出了这个：

#!/usr/bin/env python

from AppKit import *

source_image = "/Library/Desktop Pictures/Nature/Aurora.jpg"
final_image = "/Library/Desktop Pictures/.loginwindow.jpg"
font_name = "Arial"
font_size = 76
message = "My Message Here"

app = NSApplication.sharedApplication()  # remove some warnings

# read in an image
image = NSImage.alloc().initWithContentsOfFile_(source_image)
image.lockFocus()

# prepare some text attributes
text_attributes = NSMutableDictionary.alloc().init()
font = NSFont.fontWithName_size_(font_name, font_size)
text_attributes.setObject_forKey_(font, NSFontAttributeName)
text_attributes.setObject_forKey_(NSColor.blackColor, NSForegroundColorAttributeName)

# output our message
message_string = NSString.stringWithString_(message)
size = message_string.sizeWithAttributes_(text_attributes)
point = NSMakePoint(400, 400)
message_string.drawAtPoint_withAttributes_(point, text_attributes)

# write the file
image.unlockFocus()
bits = NSBitmapImageRep.alloc().initWithData_(image.TIFFRepresentation)
data = bits.representationUsingType_properties_(NSJPGFileType, nil)
data.writeToFile_atomically_(final_image, false)

当我运行它时，我明白了：

Traceback (most recent call last):
  File "/Users/clinton/Work/Problems/TellAtAGlance/ObviouslyTouched.py", line 24, in <module>
    message_string.drawAtPoint_withAttributes_(point, text_attributes)
ValueError: NSInvalidArgumentException - Class OC_PythonObject: no such selector: set

查看drawAtPoint：withAttributes：的文档，它说：“只有在NSView具有焦点时才应该调用此方法。” NSImage不是NSView的子类，但是我希望这会起作用，而且一些非常相似的东西似乎在Ruby示例中有效。

我需要做些什么来改变这项工作？

我重写了代码，将它忠实地，逐行地转换为Objective-C Foundation工具。它没有问题。 [如果有理由这样做，我会很乐意发布。]

然后问题变成了：怎么做：

[message_string drawAtPoint:point withAttributes:text_attributes];

与

不同

message_string.drawAtPoint_withAttributes_(point, text_attributes)

？有没有办法告诉哪个“OC_PythonObject”引发了NSInvalidArgumentException？

Answer 1

以上是代码中的问题：

text_attributes.setObject_forKey_(NSColor.blackColor, NSForegroundColorAttributeName)
->
text_attributes.setObject_forKey_(NSColor.blackColor(), NSForegroundColorAttributeName)

bits = NSBitmapImageRep.alloc().initWithData_(image.TIFFRepresentation)
data = bits.representationUsingType_properties_(NSJPGFileType, nil)
->
bits = NSBitmapImageRep.imageRepWithData_(image.TIFFRepresentation())
data = bits.representationUsingType_properties_(NSJPEGFileType, None)

确实是小错字。

请注意，代码的中间部分可以替换为这种更易读的变体：

# prepare some text attributes
text_attributes = { 
    NSFontAttributeName : NSFont.fontWithName_size_(font_name, font_size),
    NSForegroundColorAttributeName : NSColor.blackColor() 
}

# output our message 
NSString.drawAtPoint_withAttributes_(message, (400, 400), text_attributes)

通过查看NodeBox的源代码，psyphography.py和cocoa.py的十二行，特别是save和_getImageData方法，我了解到了这一点。

在PyObjC中在NSImage上绘制文本时出错

1 个答案: