commit 557f81ff7c010d755667f3d3eaed95f5b3dcd26f
Author: Niles Rogoff <nilesrogoff@gmail.com>
Date:   Wed Nov 30 18:54:08 2016 -0500

    Initial commit

diff --git a/libpme.py b/libpme.py
new file mode 100644
index 0000000..3d58d90
--- /dev/null
+++ b/libpme.py
@@ -0,0 +1,127 @@
+#!/usr/bin/env python3
+# PNG Metadata Editor
+# Niles Rogoff 2016
+import struct, zlib, copy
+class PME(object):
+	class color_types(object):
+		GREYSCALE = 0
+		RGB = 2
+		PALETTE = 3
+		GREYSCALE_WITH_ALPHA = 4
+		RGB_WITH_ALPHA = 6
+	def _int(self, binary):
+		return int.from_bytes(binary, byteorder="big")
+	def _bytes(self, integer, length):
+		return integer.to_bytes(length, "big")
+	def __init__(self, filename=False):
+		self._init = False
+		self._magic_number = b'\x89PNG\r\n\x1a\n'
+		self.filename = filename
+		self.chunks = []
+		if filename:
+			f = open(self.filename, "rb")
+			assert(f.read(8) == self._magic_number)
+			while True:
+				length = f.read(4)
+				label = f.read(4)
+				data = f.read(self._int(length))
+				crc = f.read(4)
+				self._verify_crc(label, data, crc)
+				self.chunks.append([length, label, data, crc])
+				if label == b"IEND":
+					break
+			assert(self.chunks[0][1] == b"IHDR")
+			self.recalculate_properties()
+		else:
+			self.chunks = [
+				[b"\0\0\0\0", b"IHDR", b"\0" * 13, b""],
+				[b"\0\0\0\0", b"IDAT", b"", b""],
+				[b"\0\0\0\0", b"IEND", b"", b""]
+			]
+			self.width = self.height = 0 # the user can decide
+			self.bit_depth = 8 # a sane default
+			self.color_type = self.color_types.RGB_WITH_ALPHA
+			self.compression_method = 0
+			self.filter_method = 0 # The only one as far as I know
+			self.interlace_method = 0
+			self._init = True
+			self.recalculate_IHDR()
+			for i in range(len(self.chunks)):
+				self.recalculate_crc(i)
+				self.recalculate_length(i)
+	# The alternative to this is to have an useless property accessor for all seven properties. Trust me, this is better.
+	def __setattr__(self, name, value):
+		super(PME, self).__setattr__(name, value)
+		if name in ["width", "height", "bit_depth", "color_type", "compression_method", "filter_method", "interlace_method"]:
+			self.recalculate_IHDR()   
+	def _calculate_crc(self, label, binary):
+		calculated = zlib.crc32(label)
+		calculated = zlib.crc32(binary, calculated) & 0xFFFFFFFF
+		return calculated
+	def _verify_crc(self, label, binary, crc):
+		assert(self._calculate_crc(label, binary) == self._int(crc))
+	def recalculate_properties(self):
+		self.width =				self._int(self.chunks[0][2][0:4])
+		self.height =				self._int(self.chunks[0][2][4:8])
+		self.bit_depth =			self._int([self.chunks[0][2][8]])
+		self.color_type =			self._int([self.chunks[0][2][9]])
+		self.compression_method =   self._int([self.chunks[0][2][10]])
+		self.filter_method =		self._int([self.chunks[0][2][11]])
+		self.interlace_method =		self._int([self.chunks[0][2][12]])
+	def recalculate_IHDR(self):
+		if not self._init: return
+		final = self._bytes(self.width, 4)
+		final += self._bytes(self.height, 4)
+		final += self._bytes(self.bit_depth, 1)
+		final += self._bytes(self.color_type, 1)
+		final += self._bytes(self.compression_method, 1)
+		final += self._bytes(self.filter_method, 1)
+		final += self._bytes(self.interlace_method, 1)
+		self.chunks[0][2] = final
+		self.recalculate_crc(0)
+		self.recalculate_length(0)
+	def _index(self, index):
+		if type(index) == list:
+			return self.chunks.index(index)
+		if type(index) == str:
+			index = bytes(index)
+		if type(index) == bytes: # PLEASE FOR THE LOVE OF GOD DO NOT DO THIS
+			index = [x[1] for x in self.chunks].index(index) # index
+		return index
+	def recalculate_crc(self, index):
+		index = self._index(index)
+		self.chunks[index][3] = self._bytes(self._calculate_crc(self.chunks[index][1], self.chunks[index][2]), 4)
+	def recalculate_length(self, index):
+		index = self._index(index)
+		self.chunks[index][0] = self._bytes(len(self.chunks[index][2]), 4)
+	def save(self,filename=False):
+		if not filename:
+			filename = self.filename
+		f = open(filename, "wb")
+		f.write(self._magic_number)
+		for chunk in self.chunks:
+			for field in chunk:
+				f.write(field)
+	decompress = zlib.decompress
+	compress = zlib.compress
+
+	def get_concatenated_idat_data(self):
+		data = b""
+		for chunk in self.chunks:
+			if chunk[1] == b'IDAT':
+				data += chunk[2]
+		return data
+	def write_raw_idat_data(self, data):
+		i = 0
+		for chunk in copy.deepcopy(self.chunks):
+			if chunk[1] == b'IDAT':
+				i += 1
+				if i > 1:
+					del self.chunks[self.chunks.index(chunk)]
+		for index in range(len(self.chunks)):
+			if self.chunks[index][1] == b'IDAT':
+				self.chunks[index][2] = data
+				self.recalculate_crc(index)
+				self.recalculate_length(index)
+				return True
+		return False