#!/usr/bin/python

import re, sys

OOPS = (
    '[', '{', '(', '<',
)

EOPS = {
    '[': ']',
    '{': '}',
    '(': ')',
    '<': '>',
}

AOPS = (
    '/', '[', ']', '{', '}', '(', ')', '<', '>',
)

EXTOPS = (
    'pdfmark',
)

OPLABEL = 'operator'

white_space = re.compile('\s')

eps_begin = re.compile('%!PS-Adobe-[0-9]*.[0-9]* EPSF-[0-9]*.[0-9]*.*')

font_type1_begin = re.compile("%!PS-AdobeFont-[0-9]+.[0-9]+[^\n\r]*")

font_rename_alias = re.compile("Type[0-9]_AH[0-9]{4}")

font_rename_name = re.compile("AH[0-9]{4}-[\S]")


class psnode:

	def __init__(self, token, parent=None):

		self.children = []
		self.token = token # the token associated with this PS node
		self.parent = parent # pointer to the parent node


class pstree:

	def __init__(self, tokens):

		self.curr_node = psnode(None)
		self.tokens = tokens

	def _find_nodes(self):

		match_name = None
		child_of_last_node = False

		for tok in self.tokens:

			print tok.name

			if match_name == tok.name:

				print tok.name, 'matched previous operator'

				match_name = None

				self.curr_node.children.append(psnode(tok, self.curr_node))

				print 'one up'

				self.curr_node = self.curr_node.parent

			elif tok.name in EXTOPS:

				print tok.name, 'matched EXTOPS'

				match_name = tok.name

				self.curr_node.children.append(psnode(tok, self.curr_node))

				print 'one down'

				self.curr_node = self.curr_node.children[-1]

			elif child_of_last_node:

				try:
					self.curr_node.children[-1].children.append(psnode(tok, self.curr_node.children[-1]))
				except IndexError:
					self.curr_node.children.append(psnode(tok, self.curr_node))

				child_of_last_node = False

			elif tok.name in OOPS:

				match = True

				if tok.name == '[':

					tmp = self.curr_node

					while tmp:

						tmp = tmp.parent

						if tmp == None:
							break
						elif tmp.token == None:
							break
						elif tmp.token.name in EXTOPS:
							print 'stay here'
							match = False
							break

				self.curr_node.children.append(psnode(tok, self.curr_node))

				if match:
					print 'one down'
					self.curr_node = self.curr_node.children[-1]

			elif (self.curr_node.token and
                              (self.curr_node.token.name in OOPS) and
                              (tok.name == EOPS[self.curr_node.token.name])):

				self.curr_node.children.append(psnode(tok, self.curr_node))
				print 'one up'
				self.curr_node = self.curr_node.parent

			elif tok.name == '/':
				self.curr_node.children.append(psnode(tok, self.curr_node))
				child_of_last_node = True
			else:
				self.curr_node.children.append(psnode(tok, self.curr_node))

	def strip_font_data(self):

		output = []
		soutput = ''
		t2fonts = {}
		drop = False
		dropuntil = None
		dropnext = False
		dropthis = False
		font = None
		checknextname = False

		for tok in self.tokens:

			# check to see if we're gonna be getting binary data and tell the tokenizer that it needs to be ready
			if tok.name == 'eexec':
				self.tokens.mode = 'type1'

			# predrop/append changes

			if checknextname and (tok.data_type == 'name'):

				if font_rename_name.match(tok.name):
					font = re.sub("AH[0-9]{4}-","",tok.name)

				if tok.name == 'if':
					dropthis = True					

				if checknextname <= 0:
					checknextname = False
				else:
					checknextname -= 1

			if tok.name == 'CIDFontName':
				checknextname = 1

			if font_rename_alias.match(tok.name) and font:
				t2fonts[tok.name] = font
				font = False

			if font_rename_alias.match(tok.name) and not font:
				try:
					tok.name = t2fonts[tok.name]


					sys.stderr.write('Warning: ' + tok.name + ' was not replaced. Please manually replace it if the output is not correct\n')

			if tok.name == 'CIDFontType':
				output.pop(-1)
				output.pop(-1)
				output.pop(-1)
				output.pop(-1)
				drop = True
				dropuntil = 'composefont'

			if font_type1_begin.match(tok.name):
				drop = True
				dropuntil = 'cleartomark'

			# drop and set/reset mode variables
			if not drop and not dropnext and not dropthis:
				output.append(tok.name)

			if dropnext:
				dropnext = False

			if dropthis:
				dropthis = False

			if tok.name == dropuntil:
				drop = False
				dropuntil = None
				if tok.name == 'cleartomark':
					checknextname = 2

			# post drop/apend changes
			if tok.name == 'composefont':
				dropnext = True

		nonewlines = 0

		for x in output:

			if str(x) in ('/'):
				soutput += str(x)
			elif str(x) == '(':
				soutput += str(x)
				nonewlines += 1
			elif nonewlines > 0 and str(x) != ')':
				soutput += str(x)+' '
			elif str(x) == ')':
				if soutput[-1] == ' ':
					soutput = soutput[:-1]+str(x)+' '
				else:
					soutput += str(x)+' '
				nonewlines -= 1
			else:
				soutput += str(x)+'\n'

		return soutput

	def ps_to_tree(self):
		self._find_nodes()
		while self.curr_node.parent: # get to the top of the tree
			self.curr_node = self.curr_node.parent
		return self.curr_node