PNG Metadata Editor
This library was written for a competition between me and Ethan, who was going to write this same thing in Java. He failed.
You can initialize a PME instance by passing it a filename to an already existing file, or no arguments, like this:
img = new libpme.PME("test.png"); # or img = new libpme.PME();
If the file is not valid, it will throw an exception
These are the values that you can access or change. Most are self explanatory. Changing any of these will automatically call
recalculate_IHDR. If you really want to suppress this behavior, just back up the IHDR chunk before changing the value. The defaults for an object created with no filename are displayed in parenthesis
width (0) height (0) bit_depth (8) color_type (color_types.RGB_WITH_ALPHA (6)) compression_method (0) filter_method (0) interlace_method (0)
Here are the methods you can call on an image, in no particular order
This simply recalculates the seven properties listed above from the data currently in the IHDR chunk.
img.height = 400 img.chunks = b'\x00\x00\x00\x00\x00\x00\x00\x00\x08\x06\x00\x00\x00' # img.height remains unchanged at 400 img.recalculate_properties() # img.height is now set to 0
This takes the seven properties listed above and destructively replaces the first chunk's data and crc with updated fields
img.width = 0 # img.chunks is now set to b'\x00\x00\x00\x00\x00\x00\x00\x00\x08\x06\x00\x00\x00' img.width = 400 img.recalculate_IHDR() # img.chunks is now set to b'\x00\x00\x01\x90\x00\x00\x00\x00\x08\x06\x00\x00\x00'
This takes an index of a chunk (see below) and overwrites the crc in that chunk with a newly calculated crc based on that chunk's label and data
img.color_type = libpme.color_types.RGB img.width = img.height = 1 img.chunks = b'\x00\xFF\x00\x00' # a 1x1 red image, assuming that the second chunk is the only IDAT chunk img.recalculate_crc(1) # img.chunks is now set to b'T\xbb\xd3\xea'
This takes an index of a chunk (see below) and updates the length of that chunk with a newly calculated length based on that chunk's data
img.color_type = libpme.color_types.RGB img.width = img.height = 1 img.chunks = img.compress(b'\x00\xFF\x00\x00') # a 1x1 red image, assuming that the second chunk is the only IDAT chunk img.recalculate_length(1) # img.chunks is now set to b'\x00\x00\x00\x04'
This saves the image to the disk, overwriting a file if it was already there. If the object was created from an existing file and no arguments are passed, it will use the original file.
img.save("red1.png") img.save() # Only if the object was created from an existing file.
Concatenates the (still compressed) data of each IDAT chunk, then returns it.
img.width = 2 img.chunks.insert(2, [b'\x00\x00\x00\x03', b'IDAT', img.compress(b'\x00\xFF\x00'), b'j\xee\xb3\xd0']) # making it a 2x1 image with a red and a green pixel img.decompress(img.get_concatenated_idat_data()) # returns '\x00\xFF\x00\x00\x00\xFF\x00';
Deletes all IDAT chunks except the first one, then sets that chunk's data to the argument it was passed, and recalculates its crc and length.
img.write_raw_idat_data(img.compress(b'\x00\x00\x00\xFF\x00\x00\xFF')) # a 2x1 image with all blue pixels
Any function that takes an index can be passed either the numerical index of the chunk (so 0 for IDAT, 1 for the second chunk, -1 for the last chunk, etc..), or a list that exists in
img.chunks, or a 4-length bytes object that is equal to the label of one of the chunks in the image. If a chunk with that label appears more than once, the first one will be used
The following color types are defined
GREYSCALE = 0 RGB = 2 PALETTE = 3 GREYSCALE_WITH_ALPHA = 4 RGB_WITH_ALPHA = 6
In this example, we will write draw a red circle in a new 100x100 file
img = libpme.PME(); img.width = img.height = 100 img.color_type = libpme.color_types.RGB newdata = b'' for y in range(100): newdata += b'\x00' # to indicate that we are writing raw pixel data, not differences for x in range(100): if (x-50)**2 + (y-50)**2 < 50**2: # If we're within 50 pixels of the center of the image newdata += b'\xFF\x00\x00' # red else: newdata += b'\xFF\xFF\xFF' # white img.write_raw_idat_data(img.compress(newdata)) img.save("output.png")
In this example, we will draw a red sine wave on a new 100x100 file
import math img = libpme.PME(); img.width = img.height = 100 img.color_type = libpme.color_types.RGB newdata = b'' for y in range(100): newdata += b'\x00' for x in range(100): if abs(math.sin(float(x) / 5) * 50 + 50 - y) < 5: newdata += b'\xFF\x00\x00' # red else: newdata += b'\xFF\xFF\xFF' # white img.write_raw_idat_data(img.compress(newdata)) img.save("output2.png")
You can now do
libpme.compress instead of having to write
Added the damaged argument to the constructor. If damaged is set to true, it will allow you to open potentially damaged png files, which maybe don't have IDAT as the first chunk, or have invalid CRCs