#-*- coding: UTF-8

from operator import itemgetter


def coldata_strip(s):
	s = s.strip()
	return s


def coldata_split(s):
	s = coldata_strip(s)
	return list(filter(bool, list(map(str.strip, s.split(' ')))))


def weight2tuple(w):
	return tuple([ int(x, base=16) for x in w.split('.')  ])


def collect_contractions(codepoints_file, contractions_file):
	'''collect codepoints and contractions from file (title is misleading)
	in form of ([ "0001", "0002"], weight)'''

	def reweight_collection(collection):
		collection = sorted(collection, key=itemgetter(1))
		count = 0  # actually starts from weight 1, weight 0 is a special case for U+0000
		prev_weight = None
		for i, (codepoint, weight) in enumerate(collection):
			if weight != prev_weight or prev_weight is None:
				count += 1
			prev_weight = weight
			collection[i] = (codepoint, count)

		return collection  # already sorted

	def split_collection(collection):
		codepoints = []
		contractions = []
		for cps, weight in collection:
			if len(cps) > 1:
				contractions.append((cps, weight))
			else:
				codepoints.append((cps, weight))

		return codepoints, contractions

	combined = []
	for filename in (codepoints_file, contractions_file, ):
		for line in open(filename, 'rt'):
			tokens = coldata_split(line)
			points, weight = points2points(tokens[:-1]), weight2tuple(tokens[-1])
			combined.append((points, weight))

	combined = reweight_collection(combined)  # assign sequential weights
	codepoints, contractions = split_collection(combined)  # already sorted

	return codepoints, contractions


def points2points(points):
	'''0000 to 000000 i.e. any-byte codepoints to 6-byte codepoints
	this will make codepoints string representation usable w/o %06X formatting'''
	return ['%06X' % (int(point, base=16)) for point in points]


def find_weight(codepoints, collection):
	'''lookup weight in list of (codepoints, weight)
	works on both: contractions and codepoints'''
	for points, weight in collection:
		if points == codepoints:
			return weight